@alicloud/aidge20260428 5.12.2 → 5.14.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/client.d.ts +774 -630
- package/dist/client.js +863 -630
- package/dist/client.js.map +1 -1
- package/dist/models/AddDisplayKnowledgeResponseBody.d.ts +3 -0
- package/dist/models/AddDisplayKnowledgeResponseBody.js.map +1 -1
- package/dist/models/AiTryOnRequest.d.ts +65 -8
- package/dist/models/AiTryOnRequest.js +10 -0
- package/dist/models/AiTryOnRequest.js.map +1 -1
- package/dist/models/AiTryOnResponseBody.d.ts +8 -8
- package/dist/models/CategoryAttributeMatchResponseBody.d.ts +3 -0
- package/dist/models/CategoryAttributeMatchResponseBody.js.map +1 -1
- package/dist/models/DeleteDisplayKnowledgeResponseBody.d.ts +3 -0
- package/dist/models/DeleteDisplayKnowledgeResponseBody.js.map +1 -1
- package/dist/models/DocumentTranslateRequest.d.ts +10 -7
- package/dist/models/DocumentTranslateRequest.js.map +1 -1
- package/dist/models/DocumentTranslateResponseBody.d.ts +3 -3
- package/dist/models/EcomVideoRecreationRequest.d.ts +11 -17
- package/dist/models/EcomVideoRecreationRequest.js.map +1 -1
- package/dist/models/EcomVideoRecreationResponseBody.d.ts +3 -3
- package/dist/models/EcomVideoRecreationShrinkRequest.d.ts +2 -2
- package/dist/models/GeneralRephotographyDetectionResponseBody.d.ts +4 -4
- package/dist/models/GenerateProductContentRequest.d.ts +11 -13
- package/dist/models/GenerateProductContentRequest.js.map +1 -1
- package/dist/models/GenerateProductContentResponseBody.d.ts +9 -7
- package/dist/models/GenerateProductContentResponseBody.js.map +1 -1
- package/dist/models/ImageMattingAsyncRequest.d.ts +1 -1
- package/dist/models/ImageMattingAsyncResponseBody.d.ts +2 -2
- package/dist/models/ImageRecognitionAsyncRequest.d.ts +14 -11
- package/dist/models/ImageRecognitionAsyncRequest.js.map +1 -1
- package/dist/models/ImageRecognitionAsyncResponseBody.d.ts +5 -5
- package/dist/models/ImageRecognitionAsyncShrinkRequest.d.ts +14 -11
- package/dist/models/ImageRecognitionAsyncShrinkRequest.js.map +1 -1
- package/dist/models/ImageRemovalProAsyncRequest.d.ts +1 -1
- package/dist/models/ImageRemovalProAsyncResponseBody.d.ts +5 -7
- package/dist/models/ImageRemovalProAsyncResponseBody.js.map +1 -1
- package/dist/models/ImageRemovalProResponseBody.d.ts +3 -0
- package/dist/models/ImageRemovalProResponseBody.js.map +1 -1
- package/dist/models/ImageRemoveAsyncRequest.d.ts +19 -8
- package/dist/models/ImageRemoveAsyncRequest.js.map +1 -1
- package/dist/models/ImageRemoveAsyncResponseBody.d.ts +8 -6
- package/dist/models/ImageRemoveAsyncResponseBody.js.map +1 -1
- package/dist/models/ImageRemoveAsyncShrinkRequest.d.ts +19 -8
- package/dist/models/ImageRemoveAsyncShrinkRequest.js.map +1 -1
- package/dist/models/ImageTranslationPlusAsyncRequest.d.ts +11 -11
- package/dist/models/ImageTranslationPlusAsyncResponseBody.d.ts +4 -4
- package/dist/models/ImageTranslationProAsyncRequest.d.ts +5 -5
- package/dist/models/ImageTranslationProAsyncResponseBody.d.ts +1 -1
- package/dist/models/ImageTranslationStandardAsyncRequest.d.ts +1 -1
- package/dist/models/ImageTranslationStandardAsyncResponseBody.d.ts +3 -3
- package/dist/models/ImageTranslationStandardRequest.d.ts +4 -4
- package/dist/models/ImageTranslationStandardResponseBody.d.ts +10 -10
- package/dist/models/ImageUpscaleRequest.d.ts +1 -1
- package/dist/models/ImageUpscaleResponseBody.d.ts +7 -7
- package/dist/models/MaterialInspectionRequest.d.ts +4 -4
- package/dist/models/MaterialInspectionResponseBody.d.ts +9 -9
- package/dist/models/MerchandisePlacementDetectionLiteResponseBody.d.ts +15 -12
- package/dist/models/MerchandisePlacementDetectionLiteResponseBody.js.map +1 -1
- package/dist/models/MerchandisePlacementDetectionProRequest.d.ts +5 -5
- package/dist/models/MerchandisePlacementDetectionProResponseBody.d.ts +6 -6
- package/dist/models/MerchandisePlacementDetectionRequest.d.ts +7 -7
- package/dist/models/MerchandisePlacementDetectionResponseBody.d.ts +19 -19
- package/dist/models/MerchandisePriceTagDetectionRequest.d.ts +1 -1
- package/dist/models/MerchandisePriceTagDetectionResponseBody.d.ts +11 -11
- package/dist/models/MerchandiseSceneDetectionRequest.d.ts +1 -1
- package/dist/models/MerchandiseSceneDetectionResponseBody.d.ts +16 -10
- package/dist/models/MerchandiseSceneDetectionResponseBody.js.map +1 -1
- package/dist/models/MilkPowderCanBottomCodeRephotographyDetectionResponseBody.d.ts +7 -7
- package/dist/models/MilkPowderCartonCodeRephotographyDetectionResponseBody.d.ts +12 -9
- package/dist/models/MilkPowderCartonCodeRephotographyDetectionResponseBody.js.map +1 -1
- package/dist/models/PrepaidTextTranslateRequest.d.ts +8 -8
- package/dist/models/PrepaidTextTranslateResponseBody.d.ts +11 -8
- package/dist/models/PrepaidTextTranslateResponseBody.js.map +1 -1
- package/dist/models/PrepaidTextTranslateShrinkRequest.d.ts +8 -8
- package/dist/models/ProductVideoGenerationRequest.d.ts +62 -0
- package/dist/models/ProductVideoGenerationRequest.js +71 -0
- package/dist/models/ProductVideoGenerationRequest.js.map +1 -0
- package/dist/models/ProductVideoGenerationResponse.d.ts +19 -0
- package/dist/models/ProductVideoGenerationResponse.js +69 -0
- package/dist/models/ProductVideoGenerationResponse.js.map +1 -0
- package/dist/models/ProductVideoGenerationResponseBody.d.ts +86 -0
- package/dist/models/ProductVideoGenerationResponseBody.js +92 -0
- package/dist/models/ProductVideoGenerationResponseBody.js.map +1 -0
- package/dist/models/ProductVideoGenerationShrinkRequest.d.ts +62 -0
- package/dist/models/ProductVideoGenerationShrinkRequest.js +68 -0
- package/dist/models/ProductVideoGenerationShrinkRequest.js.map +1 -0
- package/dist/models/QueryDisplayKnowledgeResponseBody.d.ts +3 -0
- package/dist/models/QueryDisplayKnowledgeResponseBody.js.map +1 -1
- package/dist/models/SizeChartExtractAsyncRequest.d.ts +3 -3
- package/dist/models/SizeChartExtractAsyncResponseBody.d.ts +7 -5
- package/dist/models/SizeChartExtractAsyncResponseBody.js.map +1 -1
- package/dist/models/SizeChartExtractAsyncShrinkRequest.d.ts +3 -3
- package/dist/models/StorefrontSceneRecognitionRequest.d.ts +1 -1
- package/dist/models/StorefrontSceneRecognitionResponseBody.d.ts +9 -6
- package/dist/models/StorefrontSceneRecognitionResponseBody.js.map +1 -1
- package/dist/models/SubmitProductMatchRequest.d.ts +8 -8
- package/dist/models/SubmitProductMatchResponseBody.d.ts +7 -7
- package/dist/models/TextTranslateAsyncRequest.d.ts +9 -9
- package/dist/models/TextTranslateAsyncResponseBody.d.ts +6 -6
- package/dist/models/TextTranslateAsyncShrinkRequest.d.ts +9 -9
- package/dist/models/UpdateDisplayKnowledgeResponseBody.d.ts +3 -0
- package/dist/models/UpdateDisplayKnowledgeResponseBody.js.map +1 -1
- package/dist/models/VacancyDetectionRequest.d.ts +39 -0
- package/dist/models/VacancyDetectionRequest.js +62 -0
- package/dist/models/VacancyDetectionRequest.js.map +1 -0
- package/dist/models/VacancyDetectionResponse.d.ts +19 -0
- package/dist/models/VacancyDetectionResponse.js +69 -0
- package/dist/models/VacancyDetectionResponse.js.map +1 -0
- package/dist/models/VacancyDetectionResponseBody.d.ts +83 -0
- package/dist/models/VacancyDetectionResponseBody.js +93 -0
- package/dist/models/VacancyDetectionResponseBody.js.map +1 -0
- package/dist/models/VideoGenerationRequest.d.ts +18 -17
- package/dist/models/VideoGenerationRequest.js.map +1 -1
- package/dist/models/VideoGenerationResponseBody.d.ts +2 -2
- package/dist/models/VideoTranslationRequest.d.ts +8 -17
- package/dist/models/VideoTranslationRequest.js.map +1 -1
- package/dist/models/VideoTranslationResponseBody.d.ts +4 -6
- package/dist/models/VideoTranslationResponseBody.js.map +1 -1
- package/dist/models/VideoTranslationShrinkRequest.d.ts +8 -17
- package/dist/models/VideoTranslationShrinkRequest.js.map +1 -1
- package/dist/models/VirtualTryOnResponseBody.d.ts +7 -6
- package/dist/models/VirtualTryOnResponseBody.js.map +1 -1
- package/dist/models/model.d.ts +9 -0
- package/dist/models/model.js +23 -5
- package/dist/models/model.js.map +1 -1
- package/package.json +1 -1
- package/src/client.ts +882 -630
- package/src/models/AddDisplayKnowledgeResponseBody.ts +3 -0
- package/src/models/AiTryOnRequest.ts +75 -8
- package/src/models/AiTryOnResponseBody.ts +8 -8
- package/src/models/CategoryAttributeMatchResponseBody.ts +3 -0
- package/src/models/DeleteDisplayKnowledgeResponseBody.ts +3 -0
- package/src/models/DocumentTranslateRequest.ts +10 -7
- package/src/models/DocumentTranslateResponseBody.ts +3 -3
- package/src/models/EcomVideoRecreationRequest.ts +11 -17
- package/src/models/EcomVideoRecreationResponseBody.ts +3 -3
- package/src/models/EcomVideoRecreationShrinkRequest.ts +2 -2
- package/src/models/GeneralRephotographyDetectionResponseBody.ts +4 -4
- package/src/models/GenerateProductContentRequest.ts +11 -13
- package/src/models/GenerateProductContentResponseBody.ts +9 -7
- package/src/models/ImageMattingAsyncRequest.ts +1 -1
- package/src/models/ImageMattingAsyncResponseBody.ts +2 -2
- package/src/models/ImageRecognitionAsyncRequest.ts +14 -11
- package/src/models/ImageRecognitionAsyncResponseBody.ts +5 -5
- package/src/models/ImageRecognitionAsyncShrinkRequest.ts +14 -11
- package/src/models/ImageRemovalProAsyncRequest.ts +1 -1
- package/src/models/ImageRemovalProAsyncResponseBody.ts +5 -7
- package/src/models/ImageRemovalProResponseBody.ts +3 -0
- package/src/models/ImageRemoveAsyncRequest.ts +19 -8
- package/src/models/ImageRemoveAsyncResponseBody.ts +8 -6
- package/src/models/ImageRemoveAsyncShrinkRequest.ts +19 -8
- package/src/models/ImageTranslationPlusAsyncRequest.ts +11 -11
- package/src/models/ImageTranslationPlusAsyncResponseBody.ts +4 -4
- package/src/models/ImageTranslationProAsyncRequest.ts +5 -5
- package/src/models/ImageTranslationProAsyncResponseBody.ts +1 -1
- package/src/models/ImageTranslationStandardAsyncRequest.ts +1 -1
- package/src/models/ImageTranslationStandardAsyncResponseBody.ts +3 -3
- package/src/models/ImageTranslationStandardRequest.ts +4 -4
- package/src/models/ImageTranslationStandardResponseBody.ts +10 -10
- package/src/models/ImageUpscaleRequest.ts +1 -1
- package/src/models/ImageUpscaleResponseBody.ts +7 -7
- package/src/models/MaterialInspectionRequest.ts +4 -4
- package/src/models/MaterialInspectionResponseBody.ts +9 -9
- package/src/models/MerchandisePlacementDetectionLiteResponseBody.ts +15 -12
- package/src/models/MerchandisePlacementDetectionProRequest.ts +5 -5
- package/src/models/MerchandisePlacementDetectionProResponseBody.ts +6 -6
- package/src/models/MerchandisePlacementDetectionRequest.ts +7 -7
- package/src/models/MerchandisePlacementDetectionResponseBody.ts +19 -19
- package/src/models/MerchandisePriceTagDetectionRequest.ts +1 -1
- package/src/models/MerchandisePriceTagDetectionResponseBody.ts +11 -11
- package/src/models/MerchandiseSceneDetectionRequest.ts +1 -1
- package/src/models/MerchandiseSceneDetectionResponseBody.ts +16 -10
- package/src/models/MilkPowderCanBottomCodeRephotographyDetectionResponseBody.ts +7 -7
- package/src/models/MilkPowderCartonCodeRephotographyDetectionResponseBody.ts +12 -9
- package/src/models/PrepaidTextTranslateRequest.ts +8 -8
- package/src/models/PrepaidTextTranslateResponseBody.ts +11 -8
- package/src/models/PrepaidTextTranslateShrinkRequest.ts +8 -8
- package/src/models/ProductVideoGenerationRequest.ts +88 -0
- package/src/models/ProductVideoGenerationResponse.ts +40 -0
- package/src/models/ProductVideoGenerationResponseBody.ts +124 -0
- package/src/models/ProductVideoGenerationShrinkRequest.ts +85 -0
- package/src/models/QueryDisplayKnowledgeResponseBody.ts +3 -0
- package/src/models/SizeChartExtractAsyncRequest.ts +3 -3
- package/src/models/SizeChartExtractAsyncResponseBody.ts +7 -5
- package/src/models/SizeChartExtractAsyncShrinkRequest.ts +3 -3
- package/src/models/StorefrontSceneRecognitionRequest.ts +1 -1
- package/src/models/StorefrontSceneRecognitionResponseBody.ts +9 -6
- package/src/models/SubmitProductMatchRequest.ts +8 -8
- package/src/models/SubmitProductMatchResponseBody.ts +7 -7
- package/src/models/TextTranslateAsyncRequest.ts +9 -9
- package/src/models/TextTranslateAsyncResponseBody.ts +6 -6
- package/src/models/TextTranslateAsyncShrinkRequest.ts +9 -9
- package/src/models/UpdateDisplayKnowledgeResponseBody.ts +3 -0
- package/src/models/VacancyDetectionRequest.ts +56 -0
- package/src/models/VacancyDetectionResponse.ts +40 -0
- package/src/models/VacancyDetectionResponseBody.ts +120 -0
- package/src/models/VideoGenerationRequest.ts +18 -17
- package/src/models/VideoGenerationResponseBody.ts +2 -2
- package/src/models/VideoTranslationRequest.ts +8 -17
- package/src/models/VideoTranslationResponseBody.ts +4 -6
- package/src/models/VideoTranslationShrinkRequest.ts +8 -17
- package/src/models/VirtualTryOnResponseBody.ts +7 -6
- package/src/models/model.ts +9 -0
package/dist/client.js
CHANGED
|
@@ -113,38 +113,68 @@ class Client extends openapi_core_1.default {
|
|
|
113
113
|
return await this.addDisplayKnowledgeWithOptions(request, runtime);
|
|
114
114
|
}
|
|
115
115
|
/**
|
|
116
|
-
* Generates
|
|
116
|
+
* Generates virtual try-on images by using AI algorithms to replace a model\\"s clothing with a specified garment based on the input garment and model images. The operation preserves the model\\"s posture, facial features, and other visual details while ensuring a realistic and well-fitted appearance of the garment. It returns a processed image based on the provided URL of the original product image.
|
|
117
117
|
*
|
|
118
118
|
* @remarks
|
|
119
|
-
*
|
|
120
|
-
*
|
|
121
|
-
*
|
|
122
|
-
*
|
|
123
|
-
*
|
|
124
|
-
*
|
|
125
|
-
*
|
|
126
|
-
*
|
|
127
|
-
*
|
|
128
|
-
*
|
|
129
|
-
*
|
|
130
|
-
*
|
|
131
|
-
*
|
|
132
|
-
* -
|
|
133
|
-
*
|
|
134
|
-
*
|
|
135
|
-
*
|
|
136
|
-
*
|
|
137
|
-
*
|
|
138
|
-
*
|
|
139
|
-
*
|
|
140
|
-
* | **Type** | **Model image** | **Clothing image** | **Try-on result image** |
|
|
119
|
+
* # Product overview
|
|
120
|
+
* Virtual try-on is an AI-based image generation API service that replaces the clothing on a model with specified garments based on user-provided clothing and model images. It preserves the model\\"s other details and scene details without changing the model\\"s pose or facial features, ensuring a realistic and well-fitted appearance of the garments. You only need to provide the URLs of the original product images to obtain the processed images.
|
|
121
|
+
* # Scenarios
|
|
122
|
+
* * **Virtual try-on for clothing, shoes, and hats:**
|
|
123
|
+
*
|
|
124
|
+
* Supports virtual try-on for 14 categories of clothing, shoes, hats, and accessories, including tops, coats, bottoms, dresses, jumpsuits, socks, shoes, hats, earrings, necklaces, bracelets, scarves, bags, and glasses. After you upload a person image and product images, the system intelligently matches the position and proportion of the products based on the person\\"s pose and body parts to generate natural and realistic try-on effects. This intuitively displays the on-body performance of the clothing and presents a coordinated and realistic wearing effect.
|
|
125
|
+
* * **Model replacement (model skin/face replacement)**
|
|
126
|
+
*
|
|
127
|
+
* You provide the original model scene image and the face or reference image of the target model to apply the target person\\"s face to the model in the original image. This is applicable to customer-specified models, unified brand images, and character image adaptation for different markets. It also reduces the duplication of using the same model image as other merchants.
|
|
128
|
+
*
|
|
129
|
+
* * **Pose replacement (pose fission)**
|
|
130
|
+
*
|
|
131
|
+
* You provide the original model scene image and a pose reference image with a specified pose to generate the effect of the original person and outfit in the new pose. This displays the fit, silhouette, and on-body performance of the products in different poses such as standing, raising hands, and turning around.
|
|
132
|
+
* * **E-commerce product display and marketing:**
|
|
133
|
+
*
|
|
134
|
+
* Helps merchants quickly generate model try-on images for different products, reducing the costs of traditional shooting, outfit changes, and post-production, and improving the production efficiency of product images.
|
|
135
|
+
* * **Online shopping experience optimization:**
|
|
136
|
+
*
|
|
137
|
+
* Consumers can preview the wearing effect of products through virtual try-on to more intuitively understand the style, matching, and overall effect of the products, assisting in purchase decisions and enhancing the shopping experience.
|
|
138
|
+
* # Features
|
|
139
|
+
* | Feature | Supported scope | User input and specification method | Scenario value |
|
|
141
140
|
* | --- | --- | --- | --- |
|
|
142
|
-
* |
|
|
143
|
-
*
|
|
144
|
-
*
|
|
145
|
-
*
|
|
146
|
-
*
|
|
147
|
-
*
|
|
141
|
+
* | **Clothing replacement/product wearing** | * Clothing: tops, bottoms (pants, skirts), coats, dresses, jumpsuits/rompers.
|
|
142
|
+
*
|
|
143
|
+
* * Shoes and hats: shoes, socks, hats.
|
|
144
|
+
*
|
|
145
|
+
* * Accessories: earrings, necklaces, bracelets, scarves, bags, glasses. | Upload a model image and product images, and the products will naturally fit the model. The system automatically matches the wearing position based on the clothing category, ensuring that the pose, proportion, and wearing layers are consistent with real human effects, while maintaining the person\\"s scene and pose in the original image.
|
|
146
|
+
* Try-on modes:
|
|
147
|
+
* * Supports single-item try-on.
|
|
148
|
+
*
|
|
149
|
+
* * Supports multi-item outfit matching (multiple items in one image or multiple items across multiple images).
|
|
150
|
+
*
|
|
151
|
+
* Note: The multi-item outfit matching capability was newly launched in November. | Uniformly generates the wearing effects of clothing, shoes, hats, and accessories, reducing item-by-item outfit changes and real-life shooting. |
|
|
152
|
+
* | **Model replacement/model skin replacement (face replacement)** | Supports user-specified target models.
|
|
153
|
+
* The original image can be a single-item try-on image or a multi-item outfit image. | Upload the original model scene image and provide one face photo or clear model reference image of the target person to replace the face and identity features of the model in the scene with those of the target person.
|
|
154
|
+
* The original outfit, pose, and scene are all preserved, and the generated result uses the person\\"s pose and background from the scene image. | Uses customer-specified or authorized model images to reduce duplication with existing model images and unify brand visuals. |
|
|
155
|
+
* | **Pose replacement/pose fission** | Generates images based on the user-specified pose.
|
|
156
|
+
* The original image can be a single-item try-on image or a multi-item outfit image. | Upload the original model scene image and one pose reference image to change the model\\"s pose to the action in the reference image. The person and outfit remain unchanged, and the wrinkles, drape, and layers of the clothing naturally adjust to the new pose. | Displays the fit, silhouette, and on-body style of the products in different poses, expanding the detail pages and marketing materials. |
|
|
157
|
+
* # 4. Effect examples
|
|
158
|
+
* | Scenario | Type/outfit description | Original model scene image | Product image/reference image | Effect image |
|
|
159
|
+
* | --- | --- | --- | --- | --- |
|
|
160
|
+
* | Single-item try-on | **Vest**
|
|
161
|
+
* Black buttoned sleeveless vest |  |  |  |
|
|
162
|
+
* | Single-item try-on | Top |  |  |  |
|
|
163
|
+
* | Single-item try-on | Bottom |  |  |  |
|
|
164
|
+
* | Single-item try-on | Shoes |  |  |  |
|
|
165
|
+
* | Single-item try-on | Jumpsuit |  |  |  |
|
|
166
|
+
* | Single-item try-on | Skirt |  |  |  |
|
|
167
|
+
* | Single-item try-on | Hat |  |  |  |
|
|
168
|
+
* | Multi-item try-on | **Two-item outfit**
|
|
169
|
+
* Dress + high heels |  |  |  |
|
|
170
|
+
* | Multi-item try-on | **Three-item outfit**
|
|
171
|
+
* Striped polo shirt + denim shorts + work boots |  |  |  |
|
|
172
|
+
* | Multi-item try-on | **Four-item outfit**
|
|
173
|
+
* Printed T-shirt + pants + hat + bag |  |  |  |
|
|
174
|
+
* | Model replacement | **Model skin/face replacement**
|
|
175
|
+
* Preserves the original top, pants, and scene |  |  |  |
|
|
176
|
+
* | Pose replacement | **Specified pose**
|
|
177
|
+
* Preserves the original outfit and adjusts to the hand-raising pose in the reference image |  |  |  |
|
|
148
178
|
*
|
|
149
179
|
* @param request - AiTryOnRequest
|
|
150
180
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -159,12 +189,27 @@ class Client extends openapi_core_1.default {
|
|
|
159
189
|
if (!$dara.isNull(request.clothType)) {
|
|
160
190
|
query["ClothType"] = request.clothType;
|
|
161
191
|
}
|
|
192
|
+
if (!$dara.isNull(request.mode)) {
|
|
193
|
+
query["Mode"] = request.mode;
|
|
194
|
+
}
|
|
162
195
|
if (!$dara.isNull(request.modelImageUrl)) {
|
|
163
196
|
query["ModelImageUrl"] = request.modelImageUrl;
|
|
164
197
|
}
|
|
198
|
+
if (!$dara.isNull(request.modelTemplateUrl)) {
|
|
199
|
+
query["ModelTemplateUrl"] = request.modelTemplateUrl;
|
|
200
|
+
}
|
|
201
|
+
if (!$dara.isNull(request.poseTemplateUrl)) {
|
|
202
|
+
query["PoseTemplateUrl"] = request.poseTemplateUrl;
|
|
203
|
+
}
|
|
204
|
+
if (!$dara.isNull(request.ratio)) {
|
|
205
|
+
query["Ratio"] = request.ratio;
|
|
206
|
+
}
|
|
165
207
|
if (!$dara.isNull(request.resolution)) {
|
|
166
208
|
query["Resolution"] = request.resolution;
|
|
167
209
|
}
|
|
210
|
+
if (!$dara.isNull(request.tryonType)) {
|
|
211
|
+
query["TryonType"] = request.tryonType;
|
|
212
|
+
}
|
|
168
213
|
let req = new openapi_core_2.$OpenApiUtil.OpenApiRequest({
|
|
169
214
|
query: openapi_core_2.OpenApiUtil.query(query),
|
|
170
215
|
});
|
|
@@ -182,38 +227,68 @@ class Client extends openapi_core_1.default {
|
|
|
182
227
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.AiTryOnResponse({}));
|
|
183
228
|
}
|
|
184
229
|
/**
|
|
185
|
-
* Generates
|
|
230
|
+
* Generates virtual try-on images by using AI algorithms to replace a model\\"s clothing with a specified garment based on the input garment and model images. The operation preserves the model\\"s posture, facial features, and other visual details while ensuring a realistic and well-fitted appearance of the garment. It returns a processed image based on the provided URL of the original product image.
|
|
186
231
|
*
|
|
187
232
|
* @remarks
|
|
188
|
-
*
|
|
189
|
-
*
|
|
190
|
-
*
|
|
191
|
-
*
|
|
192
|
-
*
|
|
193
|
-
*
|
|
194
|
-
*
|
|
195
|
-
*
|
|
196
|
-
*
|
|
197
|
-
*
|
|
198
|
-
*
|
|
199
|
-
*
|
|
200
|
-
*
|
|
201
|
-
* -
|
|
202
|
-
*
|
|
203
|
-
*
|
|
204
|
-
*
|
|
205
|
-
*
|
|
206
|
-
*
|
|
207
|
-
*
|
|
208
|
-
*
|
|
209
|
-
* | **Type** | **Model image** | **Clothing image** | **Try-on result image** |
|
|
233
|
+
* # Product overview
|
|
234
|
+
* Virtual try-on is an AI-based image generation API service that replaces the clothing on a model with specified garments based on user-provided clothing and model images. It preserves the model\\"s other details and scene details without changing the model\\"s pose or facial features, ensuring a realistic and well-fitted appearance of the garments. You only need to provide the URLs of the original product images to obtain the processed images.
|
|
235
|
+
* # Scenarios
|
|
236
|
+
* * **Virtual try-on for clothing, shoes, and hats:**
|
|
237
|
+
*
|
|
238
|
+
* Supports virtual try-on for 14 categories of clothing, shoes, hats, and accessories, including tops, coats, bottoms, dresses, jumpsuits, socks, shoes, hats, earrings, necklaces, bracelets, scarves, bags, and glasses. After you upload a person image and product images, the system intelligently matches the position and proportion of the products based on the person\\"s pose and body parts to generate natural and realistic try-on effects. This intuitively displays the on-body performance of the clothing and presents a coordinated and realistic wearing effect.
|
|
239
|
+
* * **Model replacement (model skin/face replacement)**
|
|
240
|
+
*
|
|
241
|
+
* You provide the original model scene image and the face or reference image of the target model to apply the target person\\"s face to the model in the original image. This is applicable to customer-specified models, unified brand images, and character image adaptation for different markets. It also reduces the duplication of using the same model image as other merchants.
|
|
242
|
+
*
|
|
243
|
+
* * **Pose replacement (pose fission)**
|
|
244
|
+
*
|
|
245
|
+
* You provide the original model scene image and a pose reference image with a specified pose to generate the effect of the original person and outfit in the new pose. This displays the fit, silhouette, and on-body performance of the products in different poses such as standing, raising hands, and turning around.
|
|
246
|
+
* * **E-commerce product display and marketing:**
|
|
247
|
+
*
|
|
248
|
+
* Helps merchants quickly generate model try-on images for different products, reducing the costs of traditional shooting, outfit changes, and post-production, and improving the production efficiency of product images.
|
|
249
|
+
* * **Online shopping experience optimization:**
|
|
250
|
+
*
|
|
251
|
+
* Consumers can preview the wearing effect of products through virtual try-on to more intuitively understand the style, matching, and overall effect of the products, assisting in purchase decisions and enhancing the shopping experience.
|
|
252
|
+
* # Features
|
|
253
|
+
* | Feature | Supported scope | User input and specification method | Scenario value |
|
|
210
254
|
* | --- | --- | --- | --- |
|
|
211
|
-
* |
|
|
212
|
-
*
|
|
213
|
-
*
|
|
214
|
-
*
|
|
215
|
-
*
|
|
216
|
-
*
|
|
255
|
+
* | **Clothing replacement/product wearing** | * Clothing: tops, bottoms (pants, skirts), coats, dresses, jumpsuits/rompers.
|
|
256
|
+
*
|
|
257
|
+
* * Shoes and hats: shoes, socks, hats.
|
|
258
|
+
*
|
|
259
|
+
* * Accessories: earrings, necklaces, bracelets, scarves, bags, glasses. | Upload a model image and product images, and the products will naturally fit the model. The system automatically matches the wearing position based on the clothing category, ensuring that the pose, proportion, and wearing layers are consistent with real human effects, while maintaining the person\\"s scene and pose in the original image.
|
|
260
|
+
* Try-on modes:
|
|
261
|
+
* * Supports single-item try-on.
|
|
262
|
+
*
|
|
263
|
+
* * Supports multi-item outfit matching (multiple items in one image or multiple items across multiple images).
|
|
264
|
+
*
|
|
265
|
+
* Note: The multi-item outfit matching capability was newly launched in November. | Uniformly generates the wearing effects of clothing, shoes, hats, and accessories, reducing item-by-item outfit changes and real-life shooting. |
|
|
266
|
+
* | **Model replacement/model skin replacement (face replacement)** | Supports user-specified target models.
|
|
267
|
+
* The original image can be a single-item try-on image or a multi-item outfit image. | Upload the original model scene image and provide one face photo or clear model reference image of the target person to replace the face and identity features of the model in the scene with those of the target person.
|
|
268
|
+
* The original outfit, pose, and scene are all preserved, and the generated result uses the person\\"s pose and background from the scene image. | Uses customer-specified or authorized model images to reduce duplication with existing model images and unify brand visuals. |
|
|
269
|
+
* | **Pose replacement/pose fission** | Generates images based on the user-specified pose.
|
|
270
|
+
* The original image can be a single-item try-on image or a multi-item outfit image. | Upload the original model scene image and one pose reference image to change the model\\"s pose to the action in the reference image. The person and outfit remain unchanged, and the wrinkles, drape, and layers of the clothing naturally adjust to the new pose. | Displays the fit, silhouette, and on-body style of the products in different poses, expanding the detail pages and marketing materials. |
|
|
271
|
+
* # 4. Effect examples
|
|
272
|
+
* | Scenario | Type/outfit description | Original model scene image | Product image/reference image | Effect image |
|
|
273
|
+
* | --- | --- | --- | --- | --- |
|
|
274
|
+
* | Single-item try-on | **Vest**
|
|
275
|
+
* Black buttoned sleeveless vest |  |  |  |
|
|
276
|
+
* | Single-item try-on | Top |  |  |  |
|
|
277
|
+
* | Single-item try-on | Bottom |  |  |  |
|
|
278
|
+
* | Single-item try-on | Shoes |  |  |  |
|
|
279
|
+
* | Single-item try-on | Jumpsuit |  |  |  |
|
|
280
|
+
* | Single-item try-on | Skirt |  |  |  |
|
|
281
|
+
* | Single-item try-on | Hat |  |  |  |
|
|
282
|
+
* | Multi-item try-on | **Two-item outfit**
|
|
283
|
+
* Dress + high heels |  |  |  |
|
|
284
|
+
* | Multi-item try-on | **Three-item outfit**
|
|
285
|
+
* Striped polo shirt + denim shorts + work boots |  |  |  |
|
|
286
|
+
* | Multi-item try-on | **Four-item outfit**
|
|
287
|
+
* Printed T-shirt + pants + hat + bag |  |  |  |
|
|
288
|
+
* | Model replacement | **Model skin/face replacement**
|
|
289
|
+
* Preserves the original top, pants, and scene |  |  |  |
|
|
290
|
+
* | Pose replacement | **Specified pose**
|
|
291
|
+
* Preserves the original outfit and adjusts to the hand-raising pose in the reference image |  |  |  |
|
|
217
292
|
*
|
|
218
293
|
* @param request - AiTryOnRequest
|
|
219
294
|
* @returns AiTryOnResponse
|
|
@@ -781,18 +856,18 @@ class Client extends openapi_core_1.default {
|
|
|
781
856
|
return await this.diduiAreaDeductionWithOptions(request, runtime);
|
|
782
857
|
}
|
|
783
858
|
/**
|
|
784
|
-
* Analyzes the area of floor displays based on product recognition, SKU asset knowledge bases, and dedicated area reasoning model building. The product first locates products in
|
|
859
|
+
* Analyzes the area of floor displays based on product recognition, SKU asset knowledge bases, and dedicated area reasoning model building. The product first locates products in floor displays, reads the actual length, width, height, and reference images of the corresponding SKUs, and infers the two side lengths and area of the bounding rectangle of the floor display area. This feature is suitable for evaluating the scale of floor displays in stores and conducting inspection reviews.
|
|
785
860
|
*
|
|
786
861
|
* @remarks
|
|
787
862
|
* *Scenarios**
|
|
788
863
|
* - Floor display area evaluation: Infers the area of the bounding rectangle based on the overall photo of the floor display and the actual dimensions of the products.
|
|
789
|
-
* - Multi-SKU mixed layout analysis: Uses the dimensions of different products to infer
|
|
790
|
-
* - Area result review: Verifies the results
|
|
864
|
+
* - Multi-SKU mixed layout analysis: Uses the dimensions of different products to infer side lengths.
|
|
865
|
+
* - Area result review: Verifies the results based on the original image, product assets, and the returned two side lengths and area.
|
|
791
866
|
* **Features**
|
|
792
867
|
* - SKU asset-driven: Specifies the asset knowledge base by using ragId. Each participating product must have a name, positive length, width, and height, and at least one HTTPS reference image.
|
|
793
868
|
* - Phased area inference: Sequentially forms decision edges, SKU projections, dimension accumulation, and the final area.
|
|
794
|
-
* - Unified units: SKU asset dimensions are stored in centimeters. The
|
|
795
|
-
* - Completeness and
|
|
869
|
+
* - Unified units: SKU asset dimensions are stored in centimeters. The result side lengths are output in meters and the area is output in square meters.
|
|
870
|
+
* - Completeness and value verification: Returns a failure if assets are missing, no valid bounding boxes are found, or the area does not match the product of the side lengths.
|
|
796
871
|
*
|
|
797
872
|
* @param request - DiduiAreaDeductionProRequest
|
|
798
873
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -824,18 +899,18 @@ class Client extends openapi_core_1.default {
|
|
|
824
899
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.DiduiAreaDeductionProResponse({}));
|
|
825
900
|
}
|
|
826
901
|
/**
|
|
827
|
-
* Analyzes the area of floor displays based on product recognition, SKU asset knowledge bases, and dedicated area reasoning model building. The product first locates products in
|
|
902
|
+
* Analyzes the area of floor displays based on product recognition, SKU asset knowledge bases, and dedicated area reasoning model building. The product first locates products in floor displays, reads the actual length, width, height, and reference images of the corresponding SKUs, and infers the two side lengths and area of the bounding rectangle of the floor display area. This feature is suitable for evaluating the scale of floor displays in stores and conducting inspection reviews.
|
|
828
903
|
*
|
|
829
904
|
* @remarks
|
|
830
905
|
* *Scenarios**
|
|
831
906
|
* - Floor display area evaluation: Infers the area of the bounding rectangle based on the overall photo of the floor display and the actual dimensions of the products.
|
|
832
|
-
* - Multi-SKU mixed layout analysis: Uses the dimensions of different products to infer
|
|
833
|
-
* - Area result review: Verifies the results
|
|
907
|
+
* - Multi-SKU mixed layout analysis: Uses the dimensions of different products to infer side lengths.
|
|
908
|
+
* - Area result review: Verifies the results based on the original image, product assets, and the returned two side lengths and area.
|
|
834
909
|
* **Features**
|
|
835
910
|
* - SKU asset-driven: Specifies the asset knowledge base by using ragId. Each participating product must have a name, positive length, width, and height, and at least one HTTPS reference image.
|
|
836
911
|
* - Phased area inference: Sequentially forms decision edges, SKU projections, dimension accumulation, and the final area.
|
|
837
|
-
* - Unified units: SKU asset dimensions are stored in centimeters. The
|
|
838
|
-
* - Completeness and
|
|
912
|
+
* - Unified units: SKU asset dimensions are stored in centimeters. The result side lengths are output in meters and the area is output in square meters.
|
|
913
|
+
* - Completeness and value verification: Returns a failure if assets are missing, no valid bounding boxes are found, or the area does not match the product of the side lengths.
|
|
839
914
|
*
|
|
840
915
|
* @param request - DiduiAreaDeductionProRequest
|
|
841
916
|
* @returns DiduiAreaDeductionProResponse
|
|
@@ -845,22 +920,29 @@ class Client extends openapi_core_1.default {
|
|
|
845
920
|
return await this.diduiAreaDeductionProWithOptions(request, runtime);
|
|
846
921
|
}
|
|
847
922
|
/**
|
|
848
|
-
* Translates documents
|
|
923
|
+
* Translates documents with a solution designed specifically for document translation. It supports over 150 languages and provides pivot translation capabilities with thousands of pivot combinations. It handles multi-scenario, multi-page, and highly complex document translation, but does not currently support scanned documents. It excels in the following aspects:
|
|
924
|
+
* - Content accuracy: translation accuracy, and accuracy of parameters and units.
|
|
925
|
+
* - Structural integrity: overall layout continuity, and restoration of page margins and layouts.
|
|
849
926
|
*
|
|
850
927
|
* @remarks
|
|
851
|
-
*
|
|
852
|
-
* Designed specifically for document translation, this
|
|
928
|
+
* # Product introduction
|
|
929
|
+
* Designed specifically for document translation, this service supports more than 150 languages and provides bridging translation capabilities with thousands of language combinations. It handles multi-scenario, multi-page, and highly complex document translations, but does not currently support scanned documents.
|
|
853
930
|
* It excels in the following aspects:
|
|
854
|
-
*
|
|
855
|
-
*
|
|
856
|
-
*
|
|
857
|
-
*
|
|
858
|
-
*
|
|
859
|
-
*
|
|
860
|
-
*
|
|
861
|
-
*
|
|
862
|
-
*
|
|
863
|
-
*
|
|
931
|
+
* * Content accuracy: translation accuracy, and accuracy of parameters and units.
|
|
932
|
+
*
|
|
933
|
+
* * Structural integrity: overall layout continuity, and restoration of page margins and layouts.
|
|
934
|
+
* # Scenarios
|
|
935
|
+
* Cross-border e-commerce product manuals, contracts, agreements, business proposals, qualification documents, teaching materials, courseware, and various other scenarios.
|
|
936
|
+
* # Features
|
|
937
|
+
* * Supports PDF, Word, and Excel formats. Supports source documents that contain multiple languages. For the detailed language list, see Section 4.4.
|
|
938
|
+
*
|
|
939
|
+
* * Supports the translation of text within images in documents, as well as complex translation scenarios involving other charts and special symbols.
|
|
940
|
+
*
|
|
941
|
+
* * Supports highly accurate layout restoration after translation.
|
|
942
|
+
*
|
|
943
|
+
* * Supports the translation of a single PDF or Word document with up to **800 pages**.
|
|
944
|
+
*
|
|
945
|
+
* * Supports custom translation results, including do-not-translate (ABC-ABC), specified translation (ABC-DEF), and translation omission (ABC-null value). This is commonly used in scenarios such as brand word protection. You can implement this by passing the corresponding intervention glossary ID when calling the API to meet your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention terms. If you need more, contact the platform for assistance. For detailed instructions on translation intervention, see [Full-modal translation - Translation intervention instruction manual](https://www.alibabacloud.com/help/en/aidge/full-modal-translation-translation-intervention-instruction-manual)
|
|
864
946
|
*
|
|
865
947
|
* @param request - DocumentTranslateRequest
|
|
866
948
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -898,22 +980,29 @@ class Client extends openapi_core_1.default {
|
|
|
898
980
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.DocumentTranslateResponse({}));
|
|
899
981
|
}
|
|
900
982
|
/**
|
|
901
|
-
* Translates documents
|
|
983
|
+
* Translates documents with a solution designed specifically for document translation. It supports over 150 languages and provides pivot translation capabilities with thousands of pivot combinations. It handles multi-scenario, multi-page, and highly complex document translation, but does not currently support scanned documents. It excels in the following aspects:
|
|
984
|
+
* - Content accuracy: translation accuracy, and accuracy of parameters and units.
|
|
985
|
+
* - Structural integrity: overall layout continuity, and restoration of page margins and layouts.
|
|
902
986
|
*
|
|
903
987
|
* @remarks
|
|
904
|
-
*
|
|
905
|
-
* Designed specifically for document translation, this
|
|
988
|
+
* # Product introduction
|
|
989
|
+
* Designed specifically for document translation, this service supports more than 150 languages and provides bridging translation capabilities with thousands of language combinations. It handles multi-scenario, multi-page, and highly complex document translations, but does not currently support scanned documents.
|
|
906
990
|
* It excels in the following aspects:
|
|
907
|
-
*
|
|
908
|
-
*
|
|
909
|
-
*
|
|
910
|
-
*
|
|
911
|
-
*
|
|
912
|
-
*
|
|
913
|
-
*
|
|
914
|
-
*
|
|
915
|
-
*
|
|
916
|
-
*
|
|
991
|
+
* * Content accuracy: translation accuracy, and accuracy of parameters and units.
|
|
992
|
+
*
|
|
993
|
+
* * Structural integrity: overall layout continuity, and restoration of page margins and layouts.
|
|
994
|
+
* # Scenarios
|
|
995
|
+
* Cross-border e-commerce product manuals, contracts, agreements, business proposals, qualification documents, teaching materials, courseware, and various other scenarios.
|
|
996
|
+
* # Features
|
|
997
|
+
* * Supports PDF, Word, and Excel formats. Supports source documents that contain multiple languages. For the detailed language list, see Section 4.4.
|
|
998
|
+
*
|
|
999
|
+
* * Supports the translation of text within images in documents, as well as complex translation scenarios involving other charts and special symbols.
|
|
1000
|
+
*
|
|
1001
|
+
* * Supports highly accurate layout restoration after translation.
|
|
1002
|
+
*
|
|
1003
|
+
* * Supports the translation of a single PDF or Word document with up to **800 pages**.
|
|
1004
|
+
*
|
|
1005
|
+
* * Supports custom translation results, including do-not-translate (ABC-ABC), specified translation (ABC-DEF), and translation omission (ABC-null value). This is commonly used in scenarios such as brand word protection. You can implement this by passing the corresponding intervention glossary ID when calling the API to meet your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention terms. If you need more, contact the platform for assistance. For detailed instructions on translation intervention, see [Full-modal translation - Translation intervention instruction manual](https://www.alibabacloud.com/help/en/aidge/full-modal-translation-translation-intervention-instruction-manual)
|
|
917
1006
|
*
|
|
918
1007
|
* @param request - DocumentTranslateRequest
|
|
919
1008
|
* @returns DocumentTranslateResponse
|
|
@@ -923,45 +1012,45 @@ class Client extends openapi_core_1.default {
|
|
|
923
1012
|
return await this.documentTranslateWithOptions(request, runtime);
|
|
924
1013
|
}
|
|
925
1014
|
/**
|
|
926
|
-
*
|
|
1015
|
+
* Provides an AI video service for e-commerce content production, covering video understanding, element replacement, quality verification, and result delivery. The service focuses on product and person consistency, detail restoration, conversion expression, and scaled production. It supports targeted replacement of products or persons based on viral video templates, quickly generating multiple versions of content while preserving original actions, camera rhythms, and marketing structures to reduce shooting, editing, and adaptation costs.
|
|
927
1016
|
*
|
|
928
1017
|
* @remarks
|
|
929
1018
|
* ## **Product introduction**
|
|
930
|
-
* E-commerce video
|
|
1019
|
+
* E-commerce video variation is an AI video service for product content production, covering video understanding, element replacement, quality verification, and result delivery. The service focuses on product and person consistency, detail restoration, conversion expression, and scaled production. It supports targeted replacement of products or persons based on viral video templates, quickly generating multiple versions of content while preserving original actions, camera rhythms, and marketing structures to reduce shooting, editing, and adaptation costs.
|
|
931
1020
|
* ## **Scenarios**
|
|
932
|
-
* - **Video material
|
|
933
|
-
* - **
|
|
934
|
-
* - **Product explanation and selling point expression:** Suitable for products such as apparel, beauty and personal care, food and
|
|
935
|
-
* - **Multi-selling point and multi-version testing:**
|
|
936
|
-
* ## **
|
|
1021
|
+
* - **Video material variation:** Uses an existing e-commerce video as a template and replaces only specified products or persons in the video, preserving the original camera angles, rhythm, and marketing expression as much as possible. This is suitable for producing materials with multiple SKUs, models, regions, and versions.
|
|
1022
|
+
* - **New product launch and rapid material generation:** For centrally launched new products, this reduces manual shooting, studio setup, post-production editing, and repetitive production processes. It quickly deploys new product video materials, helping merchants shorten the launch cycle from spotting a viral video to breaking down the script and shooting new content.
|
|
1023
|
+
* - **Product explanation and selling point expression:** Suitable for products such as apparel, beauty and personal care, food and beverages, home and daily necessities, 3C digital products, and small home appliances. It can plan the explanation structure, display actions, and camera rhythm based on the selling points of the original video, and generate explanation videos tailored for the conversion of new products.
|
|
1024
|
+
* - **Multi-selling point and multi-version testing:** Quickly generates multiple versions based on different product selling points, target audiences, scripts, camera angles, and visual styles, supporting merchants in conducting content testing and iterating delivery materials.
|
|
1025
|
+
* ## **Feature overview**
|
|
937
1026
|
* ### **Core capabilities**
|
|
938
1027
|
* * **Supports two replacement modes:**
|
|
939
1028
|
*
|
|
940
|
-
* * **Product replacement (**`product_replacement`**)**: Preserves the identity, actions, scene, and sales structure
|
|
1029
|
+
* * **Product replacement (**`product_replacement`**)**: Preserves the original video\\"s host identity, actions, scene, and sales structure, and replaces the original product with the target product.
|
|
941
1030
|
*
|
|
942
|
-
* * **Person replacement (**`person_replacement`**)**: Preserves the product, actions, scene, and sales structure
|
|
1031
|
+
* * **Person replacement (**`person_replacement`**)**: Preserves the original video\\"s product, actions, scene, and sales structure, and replaces the on-screen person with the target person from reference images or descriptions.
|
|
943
1032
|
*
|
|
944
1033
|
* * **Supports two generation modes:**
|
|
945
1034
|
*
|
|
946
|
-
* * **Reference generation (**`reference`**)**: 1:1 replication. Uses the original video as a complete reference to automatically generate a new video with the same duration. It preserves the original script, storyboard, actions, scene, and
|
|
1035
|
+
* * **Reference generation (**`reference`**)**: 1:1 replication. Uses the original video as a complete reference to automatically generate a new video with the same duration as the original video. It preserves the original script, storyboard, actions, scene, and rhythm, and only performs targeted replacement of specified content such as persons or products.
|
|
947
1036
|
*
|
|
948
|
-
* * **Rebuild generation (**`rebuild`**)**: Breaks down the script, storyboard, and key elements of the original video, retains only core content such as specified persons or products, and then redesigns and generates a new video. It is not limited by the
|
|
1037
|
+
* * **Rebuild generation (**`rebuild`**)**: Breaks down the script, storyboard, and key elements of the original video, retains only core content such as specified persons or products, and then redesigns and generates a new video. It is not limited by the original video duration and supports inputting a target duration.
|
|
949
1038
|
*
|
|
950
|
-
* * **Fully automated end-to-end generation with zero prompt
|
|
1039
|
+
* * **Fully automated end-to-end generation with zero prompt threshold**: Only requires passing the reference video and target product or person images. The system automatically completes the entire process of video understanding, script planning, storyboard design, first-frame generation, and video synthesis.
|
|
951
1040
|
*
|
|
952
|
-
* * **Product understanding and information structuring**: Identifies product categories, appearance, materials, colors, structures, and key components, and combines them with
|
|
1041
|
+
* * **Product understanding and information structuring**: Identifies product categories, appearance, materials, colors, structures, and key components, and combines them with titles, attributes, and selling points provided by merchants to generate usable video expression information.
|
|
953
1042
|
*
|
|
954
1043
|
* * **Template video element replacement**: Supports replacing products or persons within the authorized scope, and defines the replacement scope and protection scope to minimize changes in non-target areas.
|
|
955
1044
|
*
|
|
956
|
-
* * **Product subject consistency**: Maintains the stability of the product\\"s shape, color, texture, pattern, packaging, and logo across different camera
|
|
1045
|
+
* * **Product subject consistency**: Maintains the stability of the product\\"s shape, color, texture, pattern, packaging, and logo as much as possible across different camera angles, perspectives, actions, and scenes.
|
|
957
1046
|
*
|
|
958
1047
|
* * **Product fidelity**: Multi-round VL quality inspection locks the product state (color, texture, structure, print) throughout the process to prevent product feature drift.
|
|
959
1048
|
*
|
|
960
|
-
* * **Selling point script and storyboard replication**:
|
|
1049
|
+
* * **Selling point script and storyboard replication**: Follows the selling point planning, video structure, camera sequence, display actions, voiceover content, and ending style of the original video.
|
|
961
1050
|
*
|
|
962
|
-
* * **Commercial model adaptation**: Supports matching corresponding model images and
|
|
1051
|
+
* * **Commercial model adaptation**: Supports matching corresponding model images and display methods based on product style, fit, target audience, and market demand.
|
|
963
1052
|
*
|
|
964
|
-
* * **Quality verification and result feedback**: Performs multi-dimensional checks on the product subject, person,
|
|
1053
|
+
* * **Quality verification and result feedback**: Performs multi-dimensional checks on the product subject, person, clothing, text, picture integrity, audio, and overall usability, and outputs result information for manual review.
|
|
965
1054
|
*
|
|
966
1055
|
* @param tmpReq - EcomVideoRecreationRequest
|
|
967
1056
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -1001,45 +1090,45 @@ class Client extends openapi_core_1.default {
|
|
|
1001
1090
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.EcomVideoRecreationResponse({}));
|
|
1002
1091
|
}
|
|
1003
1092
|
/**
|
|
1004
|
-
*
|
|
1093
|
+
* Provides an AI video service for e-commerce content production, covering video understanding, element replacement, quality verification, and result delivery. The service focuses on product and person consistency, detail restoration, conversion expression, and scaled production. It supports targeted replacement of products or persons based on viral video templates, quickly generating multiple versions of content while preserving original actions, camera rhythms, and marketing structures to reduce shooting, editing, and adaptation costs.
|
|
1005
1094
|
*
|
|
1006
1095
|
* @remarks
|
|
1007
1096
|
* ## **Product introduction**
|
|
1008
|
-
* E-commerce video
|
|
1097
|
+
* E-commerce video variation is an AI video service for product content production, covering video understanding, element replacement, quality verification, and result delivery. The service focuses on product and person consistency, detail restoration, conversion expression, and scaled production. It supports targeted replacement of products or persons based on viral video templates, quickly generating multiple versions of content while preserving original actions, camera rhythms, and marketing structures to reduce shooting, editing, and adaptation costs.
|
|
1009
1098
|
* ## **Scenarios**
|
|
1010
|
-
* - **Video material
|
|
1011
|
-
* - **
|
|
1012
|
-
* - **Product explanation and selling point expression:** Suitable for products such as apparel, beauty and personal care, food and
|
|
1013
|
-
* - **Multi-selling point and multi-version testing:**
|
|
1014
|
-
* ## **
|
|
1099
|
+
* - **Video material variation:** Uses an existing e-commerce video as a template and replaces only specified products or persons in the video, preserving the original camera angles, rhythm, and marketing expression as much as possible. This is suitable for producing materials with multiple SKUs, models, regions, and versions.
|
|
1100
|
+
* - **New product launch and rapid material generation:** For centrally launched new products, this reduces manual shooting, studio setup, post-production editing, and repetitive production processes. It quickly deploys new product video materials, helping merchants shorten the launch cycle from spotting a viral video to breaking down the script and shooting new content.
|
|
1101
|
+
* - **Product explanation and selling point expression:** Suitable for products such as apparel, beauty and personal care, food and beverages, home and daily necessities, 3C digital products, and small home appliances. It can plan the explanation structure, display actions, and camera rhythm based on the selling points of the original video, and generate explanation videos tailored for the conversion of new products.
|
|
1102
|
+
* - **Multi-selling point and multi-version testing:** Quickly generates multiple versions based on different product selling points, target audiences, scripts, camera angles, and visual styles, supporting merchants in conducting content testing and iterating delivery materials.
|
|
1103
|
+
* ## **Feature overview**
|
|
1015
1104
|
* ### **Core capabilities**
|
|
1016
1105
|
* * **Supports two replacement modes:**
|
|
1017
1106
|
*
|
|
1018
|
-
* * **Product replacement (**`product_replacement`**)**: Preserves the identity, actions, scene, and sales structure
|
|
1107
|
+
* * **Product replacement (**`product_replacement`**)**: Preserves the original video\\"s host identity, actions, scene, and sales structure, and replaces the original product with the target product.
|
|
1019
1108
|
*
|
|
1020
|
-
* * **Person replacement (**`person_replacement`**)**: Preserves the product, actions, scene, and sales structure
|
|
1109
|
+
* * **Person replacement (**`person_replacement`**)**: Preserves the original video\\"s product, actions, scene, and sales structure, and replaces the on-screen person with the target person from reference images or descriptions.
|
|
1021
1110
|
*
|
|
1022
1111
|
* * **Supports two generation modes:**
|
|
1023
1112
|
*
|
|
1024
|
-
* * **Reference generation (**`reference`**)**: 1:1 replication. Uses the original video as a complete reference to automatically generate a new video with the same duration. It preserves the original script, storyboard, actions, scene, and
|
|
1113
|
+
* * **Reference generation (**`reference`**)**: 1:1 replication. Uses the original video as a complete reference to automatically generate a new video with the same duration as the original video. It preserves the original script, storyboard, actions, scene, and rhythm, and only performs targeted replacement of specified content such as persons or products.
|
|
1025
1114
|
*
|
|
1026
|
-
* * **Rebuild generation (**`rebuild`**)**: Breaks down the script, storyboard, and key elements of the original video, retains only core content such as specified persons or products, and then redesigns and generates a new video. It is not limited by the
|
|
1115
|
+
* * **Rebuild generation (**`rebuild`**)**: Breaks down the script, storyboard, and key elements of the original video, retains only core content such as specified persons or products, and then redesigns and generates a new video. It is not limited by the original video duration and supports inputting a target duration.
|
|
1027
1116
|
*
|
|
1028
|
-
* * **Fully automated end-to-end generation with zero prompt
|
|
1117
|
+
* * **Fully automated end-to-end generation with zero prompt threshold**: Only requires passing the reference video and target product or person images. The system automatically completes the entire process of video understanding, script planning, storyboard design, first-frame generation, and video synthesis.
|
|
1029
1118
|
*
|
|
1030
|
-
* * **Product understanding and information structuring**: Identifies product categories, appearance, materials, colors, structures, and key components, and combines them with
|
|
1119
|
+
* * **Product understanding and information structuring**: Identifies product categories, appearance, materials, colors, structures, and key components, and combines them with titles, attributes, and selling points provided by merchants to generate usable video expression information.
|
|
1031
1120
|
*
|
|
1032
1121
|
* * **Template video element replacement**: Supports replacing products or persons within the authorized scope, and defines the replacement scope and protection scope to minimize changes in non-target areas.
|
|
1033
1122
|
*
|
|
1034
|
-
* * **Product subject consistency**: Maintains the stability of the product\\"s shape, color, texture, pattern, packaging, and logo across different camera
|
|
1123
|
+
* * **Product subject consistency**: Maintains the stability of the product\\"s shape, color, texture, pattern, packaging, and logo as much as possible across different camera angles, perspectives, actions, and scenes.
|
|
1035
1124
|
*
|
|
1036
1125
|
* * **Product fidelity**: Multi-round VL quality inspection locks the product state (color, texture, structure, print) throughout the process to prevent product feature drift.
|
|
1037
1126
|
*
|
|
1038
|
-
* * **Selling point script and storyboard replication**:
|
|
1127
|
+
* * **Selling point script and storyboard replication**: Follows the selling point planning, video structure, camera sequence, display actions, voiceover content, and ending style of the original video.
|
|
1039
1128
|
*
|
|
1040
|
-
* * **Commercial model adaptation**: Supports matching corresponding model images and
|
|
1129
|
+
* * **Commercial model adaptation**: Supports matching corresponding model images and display methods based on product style, fit, target audience, and market demand.
|
|
1041
1130
|
*
|
|
1042
|
-
* * **Quality verification and result feedback**: Performs multi-dimensional checks on the product subject, person,
|
|
1131
|
+
* * **Quality verification and result feedback**: Performs multi-dimensional checks on the product subject, person, clothing, text, picture integrity, audio, and overall usability, and outputs result information for manual review.
|
|
1043
1132
|
*
|
|
1044
1133
|
* @param request - EcomVideoRecreationRequest
|
|
1045
1134
|
* @returns EcomVideoRecreationResponse
|
|
@@ -1049,18 +1138,18 @@ class Client extends openapi_core_1.default {
|
|
|
1049
1138
|
return await this.ecomVideoRecreationWithOptions(request, runtime);
|
|
1050
1139
|
}
|
|
1051
1140
|
/**
|
|
1052
|
-
* Detects whether an image is recaptured and
|
|
1141
|
+
* Detects whether an image is recaptured and locates main physical products, screen display areas, printed photos, and visible recapture clues based on recapture recognition and multimodal large model element localization. This general-purpose image moderation service applies to risk screening and manual review of store inspection photos across multiple categories.
|
|
1053
1142
|
*
|
|
1054
1143
|
* @remarks
|
|
1055
1144
|
* *Scenarios**
|
|
1056
|
-
* 1. Inspection photo review: Screens
|
|
1057
|
-
* 2. Screen or printed photo recapture verification:
|
|
1145
|
+
* 1. Inspection photo review: Screens photos submitted by stores for recapture risks.
|
|
1146
|
+
* 2. Screen or printed photo recapture verification: Locates visible screens, printed photos, and other media.
|
|
1058
1147
|
* 3. Review clue display: Annotates products and visible clues on the original image to assist in manual review.
|
|
1059
1148
|
* **Features**
|
|
1060
|
-
* 1. General recapture detection: Outputs a Boolean
|
|
1149
|
+
* 1. General recapture detection: Outputs a Boolean conclusion for isFake.
|
|
1061
1150
|
* 2. Review element localization: Distinguishes physical products, screens, printed photos, and visible recapture clues. Products inside screens are not repeatedly output as physical products.
|
|
1062
1151
|
* 3. Separate presentation of clues and conclusions: Localization text only describes visible phenomena and does not directly use normal reflections or standard packaging prints as recapture evidence.
|
|
1063
|
-
* Supports
|
|
1152
|
+
* Supports single HTTPS image input. Supplementary localization returns up to 12 targets. Each target includes a category, a bounding box, and up to 24 characters of visible text, and uses coverage to indicate whether the localization is complete.
|
|
1064
1153
|
*
|
|
1065
1154
|
* @param request - GeneralRephotographyDetectionRequest
|
|
1066
1155
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -1089,18 +1178,18 @@ class Client extends openapi_core_1.default {
|
|
|
1089
1178
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.GeneralRephotographyDetectionResponse({}));
|
|
1090
1179
|
}
|
|
1091
1180
|
/**
|
|
1092
|
-
* Detects whether an image is recaptured and
|
|
1181
|
+
* Detects whether an image is recaptured and locates main physical products, screen display areas, printed photos, and visible recapture clues based on recapture recognition and multimodal large model element localization. This general-purpose image moderation service applies to risk screening and manual review of store inspection photos across multiple categories.
|
|
1093
1182
|
*
|
|
1094
1183
|
* @remarks
|
|
1095
1184
|
* *Scenarios**
|
|
1096
|
-
* 1. Inspection photo review: Screens
|
|
1097
|
-
* 2. Screen or printed photo recapture verification:
|
|
1185
|
+
* 1. Inspection photo review: Screens photos submitted by stores for recapture risks.
|
|
1186
|
+
* 2. Screen or printed photo recapture verification: Locates visible screens, printed photos, and other media.
|
|
1098
1187
|
* 3. Review clue display: Annotates products and visible clues on the original image to assist in manual review.
|
|
1099
1188
|
* **Features**
|
|
1100
|
-
* 1. General recapture detection: Outputs a Boolean
|
|
1189
|
+
* 1. General recapture detection: Outputs a Boolean conclusion for isFake.
|
|
1101
1190
|
* 2. Review element localization: Distinguishes physical products, screens, printed photos, and visible recapture clues. Products inside screens are not repeatedly output as physical products.
|
|
1102
1191
|
* 3. Separate presentation of clues and conclusions: Localization text only describes visible phenomena and does not directly use normal reflections or standard packaging prints as recapture evidence.
|
|
1103
|
-
* Supports
|
|
1192
|
+
* Supports single HTTPS image input. Supplementary localization returns up to 12 targets. Each target includes a category, a bounding box, and up to 24 characters of visible text, and uses coverage to indicate whether the localization is complete.
|
|
1104
1193
|
*
|
|
1105
1194
|
* @param request - GeneralRephotographyDetectionRequest
|
|
1106
1195
|
* @returns GeneralRephotographyDetectionResponse
|
|
@@ -1110,22 +1199,22 @@ class Client extends openapi_core_1.default {
|
|
|
1110
1199
|
return await this.generalRephotographyDetectionWithOptions(request, runtime);
|
|
1111
1200
|
}
|
|
1112
1201
|
/**
|
|
1113
|
-
* Generates or rewrites e-commerce product copy,
|
|
1202
|
+
* Generates or rewrites e-commerce product copy, such as titles, short highlights, search keywords, bulleted selling points, and product descriptions, based on input product information or existing copy, and adapts the content type and length to the requirements of the target platform. Supports Chinese and English input materials and optional multilingual translation for scenarios such as domestic product listing, cross-border product distribution, and product copy localization. Uses synchronous invocation to process one copy type for a single product per request.
|
|
1114
1203
|
*
|
|
1115
1204
|
* @remarks
|
|
1116
1205
|
* # Product introduction
|
|
1117
|
-
*
|
|
1118
|
-
*
|
|
1206
|
+
* E-commerce product copy generation is an AI text service for e-commerce product content production. By inputting product materials or existing copy, you can generate or rewrite product titles, short highlights, search keywords, bullet points, and product descriptions, and adapt the content type and length to the target platform as needed.
|
|
1207
|
+
* It supports Chinese and English material inputs and optional multilingual translation, and applies to scenarios such as domestic product listing, cross-border distribution, and product copy localization. The API uses synchronous calls and processes one copy type for one product per request.
|
|
1119
1208
|
* # Scenarios
|
|
1120
1209
|
* | **Scenario** | **Description** |
|
|
1121
1210
|
* | --- | --- |
|
|
1122
|
-
* | Same-platform localization | Select rewrite
|
|
1211
|
+
* | Same-platform localization | Select rewrite and translate as needed to retain the original facts and type structure. |
|
|
1123
1212
|
* | Cross-platform content adaptation | Select the target platform to generate or rewrite content according to its configuration limits. The source platform of the input is not required. |
|
|
1124
|
-
* | New product material
|
|
1125
|
-
* | Batch operation tool integration | The caller submits requests product by product. The workflow itself does not provide batch task or
|
|
1213
|
+
* | New product material production | Select generate to generate the target type based on product materials. Categories and attributes can be used to supplement facts. |
|
|
1214
|
+
* | Batch operation tool integration | The caller submits requests product by product. The workflow itself does not provide batch task or queue APIs. |
|
|
1126
1215
|
* # Features
|
|
1127
1216
|
* ## Core capabilities
|
|
1128
|
-
* This capability supports the
|
|
1217
|
+
* This capability supports the production of different copy content types in domestic and cross-border e-commerce scenarios. It can produce titles, detailed descriptions, keywords, highlights, and descriptions that adapt to the content rules (length requirements and writing styles) of different platforms, and supports cross-language production.
|
|
1129
1218
|
* | Content type | Platform | | | | | | Target language |
|
|
1130
1219
|
* | --- | --- | --- | --- | --- | --- | --- | --- |
|
|
1131
1220
|
* | | Amazon | Temu | TikTok | Shopify | Wildberries | Tmall |
|
|
@@ -1135,19 +1224,19 @@ class Client extends openapi_core_1.default {
|
|
|
1135
1224
|
* | **bulletPoint** | ✓ | | | ✓ | | |
|
|
1136
1225
|
* | **keywords** | ✓ | | | | | |
|
|
1137
1226
|
* | **description** | ✓ | ✓ | ✓ | ✓ | ✓ | |
|
|
1138
|
-
* Based on the
|
|
1227
|
+
* Based on the content inputs provided by users, two production modes are supported:
|
|
1139
1228
|
* | **Mode** | **Description** | **Recommended scenario** |
|
|
1140
1229
|
* | --- | --- | --- |
|
|
1141
|
-
* | Generate | Reorganizes and writes copy based on product
|
|
1142
|
-
* | Rewrite | Retains the original information and main structure
|
|
1230
|
+
* | Generate | Reorganizes and writes copy based on product materials. | Only product parameters, attributes, or selling points are available, or other types of product materials are available, and new titles, keywords, product descriptions, and other copy need to be created. |
|
|
1231
|
+
* | Rewrite | Retains the original information and main structure, and optimizes the expression. | Existing product copy is available, and you want to adjust the wording, improve readability, or localize it with translation. |
|
|
1143
1232
|
* # Examples
|
|
1144
1233
|
* The following examples use the Amazon platform:
|
|
1145
1234
|
* | **Content type** | **Source language** | **Source content** | **Target language** | **Generated result** | **Capability advantage** |
|
|
1146
1235
|
* | --- | --- | --- | --- | --- | --- |
|
|
1147
|
-
* | bulletPoint | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtersystem\\] Silikontrichter und Metallfiltergitter für 3D-Druck; Einfüllen von gebrauchtem lichtempfindlichem Harz in Flaschen mit 15 mm Halsweite | Extracts core selling points, retains key information such as bottle neck size, and presents
|
|
1236
|
+
* | bulletPoint | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtersystem\\] Silikontrichter und Metallfiltergitter für 3D-Druck; Einfüllen von gebrauchtem lichtempfindlichem Harz in Flaschen mit 15 mm Halsweite | Extracts core selling points, retains key information such as bottle neck size, and presents it in German bullet points for quick reading. |
|
|
1148
1237
|
* | description | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtrationssystem\\] Diese Kombination aus Silikontrichter und Metallfilternetz richtet sich an alle, die mit 3D-Druck arbeiten. Der flexible Silikontrichter passt sicher auf Standardflaschen mit einer Halsweite von 15 mm und ermöglicht ein sauberes Zurückfüllen von gebrauchtem lichtempfindlichem Harz ohne Kleckern. Das feine Metallfilter fängt verfestigte Partikel zuverlässig ab, sodass Ihr Harz frei von Verunreinigungen bleibt und seine Nutzbarkeit erhalten wird. Die robuste Konstruktion des Filters unterstützt eine gleichmäßige Durchflussrate beim Umfüllen. | Organizes the product purpose, usage, and filtration features into a coherent German introduction for easy understanding. |
|
|
1149
|
-
* | highlight | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtrationssystem\\] Silikontrichter, Metallfilternetz, für 3D-Druck-Harz |
|
|
1150
|
-
* | keywords | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | Silikontrichter Metallfilter Harz Filterungssystem 3D Drucker Harz umfüllen Metall Sieb für Photosensitive Harze | Extracts German search terms related to materials, categories, and purposes
|
|
1238
|
+
* | highlight | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtrationssystem\\] Silikontrichter, Metallfilternetz, für 3D-Druck-Harz | Simplifies long materials, highlights core components and purposes, and makes it easy to grasp product features at a glance. |
|
|
1239
|
+
* | keywords | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | Silikontrichter Metallfilter Harz Filterungssystem 3D Drucker Harz umfüllen Metall Sieb für Photosensitive Harze | Extracts German search terms related to materials, categories, and purposes to reduce manual sorting and translation. |
|
|
1151
1240
|
* | title | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | Silikontrichter mit Metallfilter für 3D-Druck-Harz | Extracts the product name, core components, and purpose to generate a concise and clear German title. |
|
|
1152
1241
|
*
|
|
1153
1242
|
* @param request - GenerateProductContentRequest
|
|
@@ -1210,22 +1299,22 @@ class Client extends openapi_core_1.default {
|
|
|
1210
1299
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.GenerateProductContentResponse({}));
|
|
1211
1300
|
}
|
|
1212
1301
|
/**
|
|
1213
|
-
* Generates or rewrites e-commerce product copy,
|
|
1302
|
+
* Generates or rewrites e-commerce product copy, such as titles, short highlights, search keywords, bulleted selling points, and product descriptions, based on input product information or existing copy, and adapts the content type and length to the requirements of the target platform. Supports Chinese and English input materials and optional multilingual translation for scenarios such as domestic product listing, cross-border product distribution, and product copy localization. Uses synchronous invocation to process one copy type for a single product per request.
|
|
1214
1303
|
*
|
|
1215
1304
|
* @remarks
|
|
1216
1305
|
* # Product introduction
|
|
1217
|
-
*
|
|
1218
|
-
*
|
|
1306
|
+
* E-commerce product copy generation is an AI text service for e-commerce product content production. By inputting product materials or existing copy, you can generate or rewrite product titles, short highlights, search keywords, bullet points, and product descriptions, and adapt the content type and length to the target platform as needed.
|
|
1307
|
+
* It supports Chinese and English material inputs and optional multilingual translation, and applies to scenarios such as domestic product listing, cross-border distribution, and product copy localization. The API uses synchronous calls and processes one copy type for one product per request.
|
|
1219
1308
|
* # Scenarios
|
|
1220
1309
|
* | **Scenario** | **Description** |
|
|
1221
1310
|
* | --- | --- |
|
|
1222
|
-
* | Same-platform localization | Select rewrite
|
|
1311
|
+
* | Same-platform localization | Select rewrite and translate as needed to retain the original facts and type structure. |
|
|
1223
1312
|
* | Cross-platform content adaptation | Select the target platform to generate or rewrite content according to its configuration limits. The source platform of the input is not required. |
|
|
1224
|
-
* | New product material
|
|
1225
|
-
* | Batch operation tool integration | The caller submits requests product by product. The workflow itself does not provide batch task or
|
|
1313
|
+
* | New product material production | Select generate to generate the target type based on product materials. Categories and attributes can be used to supplement facts. |
|
|
1314
|
+
* | Batch operation tool integration | The caller submits requests product by product. The workflow itself does not provide batch task or queue APIs. |
|
|
1226
1315
|
* # Features
|
|
1227
1316
|
* ## Core capabilities
|
|
1228
|
-
* This capability supports the
|
|
1317
|
+
* This capability supports the production of different copy content types in domestic and cross-border e-commerce scenarios. It can produce titles, detailed descriptions, keywords, highlights, and descriptions that adapt to the content rules (length requirements and writing styles) of different platforms, and supports cross-language production.
|
|
1229
1318
|
* | Content type | Platform | | | | | | Target language |
|
|
1230
1319
|
* | --- | --- | --- | --- | --- | --- | --- | --- |
|
|
1231
1320
|
* | | Amazon | Temu | TikTok | Shopify | Wildberries | Tmall |
|
|
@@ -1235,19 +1324,19 @@ class Client extends openapi_core_1.default {
|
|
|
1235
1324
|
* | **bulletPoint** | ✓ | | | ✓ | | |
|
|
1236
1325
|
* | **keywords** | ✓ | | | | | |
|
|
1237
1326
|
* | **description** | ✓ | ✓ | ✓ | ✓ | ✓ | |
|
|
1238
|
-
* Based on the
|
|
1327
|
+
* Based on the content inputs provided by users, two production modes are supported:
|
|
1239
1328
|
* | **Mode** | **Description** | **Recommended scenario** |
|
|
1240
1329
|
* | --- | --- | --- |
|
|
1241
|
-
* | Generate | Reorganizes and writes copy based on product
|
|
1242
|
-
* | Rewrite | Retains the original information and main structure
|
|
1330
|
+
* | Generate | Reorganizes and writes copy based on product materials. | Only product parameters, attributes, or selling points are available, or other types of product materials are available, and new titles, keywords, product descriptions, and other copy need to be created. |
|
|
1331
|
+
* | Rewrite | Retains the original information and main structure, and optimizes the expression. | Existing product copy is available, and you want to adjust the wording, improve readability, or localize it with translation. |
|
|
1243
1332
|
* # Examples
|
|
1244
1333
|
* The following examples use the Amazon platform:
|
|
1245
1334
|
* | **Content type** | **Source language** | **Source content** | **Target language** | **Generated result** | **Capability advantage** |
|
|
1246
1335
|
* | --- | --- | --- | --- | --- | --- |
|
|
1247
|
-
* | bulletPoint | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtersystem\\] Silikontrichter und Metallfiltergitter für 3D-Druck; Einfüllen von gebrauchtem lichtempfindlichem Harz in Flaschen mit 15 mm Halsweite | Extracts core selling points, retains key information such as bottle neck size, and presents
|
|
1336
|
+
* | bulletPoint | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtersystem\\] Silikontrichter und Metallfiltergitter für 3D-Druck; Einfüllen von gebrauchtem lichtempfindlichem Harz in Flaschen mit 15 mm Halsweite | Extracts core selling points, retains key information such as bottle neck size, and presents it in German bullet points for quick reading. |
|
|
1248
1337
|
* | description | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtrationssystem\\] Diese Kombination aus Silikontrichter und Metallfilternetz richtet sich an alle, die mit 3D-Druck arbeiten. Der flexible Silikontrichter passt sicher auf Standardflaschen mit einer Halsweite von 15 mm und ermöglicht ein sauberes Zurückfüllen von gebrauchtem lichtempfindlichem Harz ohne Kleckern. Das feine Metallfilter fängt verfestigte Partikel zuverlässig ab, sodass Ihr Harz frei von Verunreinigungen bleibt und seine Nutzbarkeit erhalten wird. Die robuste Konstruktion des Filters unterstützt eine gleichmäßige Durchflussrate beim Umfüllen. | Organizes the product purpose, usage, and filtration features into a coherent German introduction for easy understanding. |
|
|
1249
|
-
* | highlight | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtrationssystem\\] Silikontrichter, Metallfilternetz, für 3D-Druck-Harz |
|
|
1250
|
-
* | keywords | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | Silikontrichter Metallfilter Harz Filterungssystem 3D Drucker Harz umfüllen Metall Sieb für Photosensitive Harze | Extracts German search terms related to materials, categories, and purposes
|
|
1338
|
+
* | highlight | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | [Vielseitiges Harz-Filtrationssystem\\] Silikontrichter, Metallfilternetz, für 3D-Druck-Harz | Simplifies long materials, highlights core components and purposes, and makes it easy to grasp product features at a glance. |
|
|
1339
|
+
* | keywords | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | Silikontrichter Metallfilter Harz Filterungssystem 3D Drucker Harz umfüllen Metall Sieb für Photosensitive Harze | Extracts German search terms related to materials, categories, and purposes to reduce manual sorting and translation. |
|
|
1251
1340
|
* | title | en | [VERSATILE RESIN FILTRATION SYSTEM\\] This silicone funnel and metal filter mesh combination is designed for all 3D printing enthusiasts. It is ideal for pouring used photosensitive resin back into standard bottles with a 15mm neck width ensuring a clean transfer every time. The durable metal filter effectively catches all solidified debris protecting your valuable resin from and extending its usability. | de | Silikontrichter mit Metallfilter für 3D-Druck-Harz | Extracts the product name, core components, and purpose to generate a concise and clear German title. |
|
|
1252
1341
|
*
|
|
1253
1342
|
* @param request - GenerateProductContentRequest
|
|
@@ -1440,24 +1529,24 @@ class Client extends openapi_core_1.default {
|
|
|
1440
1529
|
return await this.imageMattingWithOptions(request, runtime);
|
|
1441
1530
|
}
|
|
1442
1531
|
/**
|
|
1443
|
-
*
|
|
1532
|
+
* Automatically identifies the salient subject in an image, separates the subject from the background, and returns the subject image with the background removed. This operation also provides multiple background options and custom size options to optimize product displays.
|
|
1444
1533
|
*
|
|
1445
1534
|
* @remarks
|
|
1446
1535
|
* # Scenarios
|
|
1447
1536
|
* * Supports secondary editing of materials: After separating the image subject from the background, you can further edit the image.
|
|
1448
1537
|
*
|
|
1449
|
-
* * Improves material processing efficiency: Quickly extracts the subjects of various material images
|
|
1538
|
+
* * Improves material processing efficiency: Quickly extracts the subjects of various material images for background replacement, improving the production efficiency of advertising images and marketing posters.
|
|
1450
1539
|
* # Features.
|
|
1451
|
-
* #### 1. One-click
|
|
1452
|
-
* * **AI-based automatic recognition**: Based on deep learning models, intelligently distinguishes foreground subjects (portraits, products, objects, etc.) from the background without manual outlining.
|
|
1540
|
+
* #### 1. One-click accurate subject recognition
|
|
1541
|
+
* * **AI-based automatic recognition**: Based on deep learning models, it intelligently distinguishes foreground subjects (portraits, products, objects, etc.) from the background without manual outlining.
|
|
1453
1542
|
*
|
|
1454
|
-
* * **Multi-scenario adaptation**: Supports
|
|
1543
|
+
* * **Multi-scenario adaptation**: Supports accurate segmentation in complex scenarios (such as hair, transparent objects, and fur edges) with an accuracy rate of over 95%.
|
|
1455
1544
|
* #### 2. Intelligent edge optimization
|
|
1456
1545
|
* * **Detail-level processing**: Performs automatic optimization on edge jaggies and semi-transparent areas (such as smoke and glass) to retain natural transition effects.
|
|
1457
1546
|
* #### 3. Multi-format output and background operations
|
|
1458
1547
|
* * **Transparent background export**: Directly generates PNG images with transparent backgrounds and white backgrounds, adapting to e-commerce and design scenarios.
|
|
1459
1548
|
*
|
|
1460
|
-
* * **
|
|
1549
|
+
* * **Custom background color support**
|
|
1461
1550
|
* # Effect comparison
|
|
1462
1551
|
* | **Original image (input)** | **Result image (output)** |
|
|
1463
1552
|
* | --- | --- |
|
|
@@ -1507,24 +1596,24 @@ class Client extends openapi_core_1.default {
|
|
|
1507
1596
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ImageMattingAsyncResponse({}));
|
|
1508
1597
|
}
|
|
1509
1598
|
/**
|
|
1510
|
-
*
|
|
1599
|
+
* Automatically identifies the salient subject in an image, separates the subject from the background, and returns the subject image with the background removed. This operation also provides multiple background options and custom size options to optimize product displays.
|
|
1511
1600
|
*
|
|
1512
1601
|
* @remarks
|
|
1513
1602
|
* # Scenarios
|
|
1514
1603
|
* * Supports secondary editing of materials: After separating the image subject from the background, you can further edit the image.
|
|
1515
1604
|
*
|
|
1516
|
-
* * Improves material processing efficiency: Quickly extracts the subjects of various material images
|
|
1605
|
+
* * Improves material processing efficiency: Quickly extracts the subjects of various material images for background replacement, improving the production efficiency of advertising images and marketing posters.
|
|
1517
1606
|
* # Features.
|
|
1518
|
-
* #### 1. One-click
|
|
1519
|
-
* * **AI-based automatic recognition**: Based on deep learning models, intelligently distinguishes foreground subjects (portraits, products, objects, etc.) from the background without manual outlining.
|
|
1607
|
+
* #### 1. One-click accurate subject recognition
|
|
1608
|
+
* * **AI-based automatic recognition**: Based on deep learning models, it intelligently distinguishes foreground subjects (portraits, products, objects, etc.) from the background without manual outlining.
|
|
1520
1609
|
*
|
|
1521
|
-
* * **Multi-scenario adaptation**: Supports
|
|
1610
|
+
* * **Multi-scenario adaptation**: Supports accurate segmentation in complex scenarios (such as hair, transparent objects, and fur edges) with an accuracy rate of over 95%.
|
|
1522
1611
|
* #### 2. Intelligent edge optimization
|
|
1523
1612
|
* * **Detail-level processing**: Performs automatic optimization on edge jaggies and semi-transparent areas (such as smoke and glass) to retain natural transition effects.
|
|
1524
1613
|
* #### 3. Multi-format output and background operations
|
|
1525
1614
|
* * **Transparent background export**: Directly generates PNG images with transparent backgrounds and white backgrounds, adapting to e-commerce and design scenarios.
|
|
1526
1615
|
*
|
|
1527
|
-
* * **
|
|
1616
|
+
* * **Custom background color support**
|
|
1528
1617
|
* # Effect comparison
|
|
1529
1618
|
* | **Original image (input)** | **Result image (output)** |
|
|
1530
1619
|
* | --- | --- |
|
|
@@ -1744,15 +1833,15 @@ class Client extends openapi_core_1.default {
|
|
|
1744
1833
|
return await this.imageRecognitionWithOptions(request, runtime);
|
|
1745
1834
|
}
|
|
1746
1835
|
/**
|
|
1747
|
-
* Identifies specific elements in e-commerce images.
|
|
1836
|
+
* Identifies specific elements in e-commerce images. This API extracts image details to help users quickly identify text, logos, watermarks, and text-containing color blocks in the main subject and background, significantly improving the efficiency and accuracy of image filtering.
|
|
1748
1837
|
*
|
|
1749
1838
|
* @remarks
|
|
1750
1839
|
* # Scenarios
|
|
1751
1840
|
* 1. Intelligent image filtering and optimization:
|
|
1752
|
-
* Merchants and platforms can use the intelligent element recognition API to filter images and automatically identify images containing specific elements such as watermarks and logos. After filtering, they can
|
|
1841
|
+
* Merchants and platforms can use the intelligent element recognition API to filter images and automatically identify images containing specific elements such as watermarks and logos. After filtering, they can call the intelligent elimination API to optimize these images, improve image quality, and shorten the product release cycle.
|
|
1753
1842
|
* 2. Automated image processing and efficiency improvement:
|
|
1754
1843
|
* Through the intelligent element recognition API, merchants and platforms can automate the image processing workflow and automatically identify and mark image elements that need to be optimized. This workflow reduces manual intervention and improves the efficiency of image processing.
|
|
1755
|
-
* #
|
|
1844
|
+
* # Feature description
|
|
1756
1845
|
* **1. Definitions of image subject, image background, text, logo, watermark, and text-containing color block: For details, see the relevant examples in the elimination documentation (reserved text link).**
|
|
1757
1846
|
* * Image subject: The core product area in the image.
|
|
1758
1847
|
*
|
|
@@ -1762,11 +1851,11 @@ class Client extends openapi_core_1.default {
|
|
|
1762
1851
|
*
|
|
1763
1852
|
* * Logo: A logo or trademark.
|
|
1764
1853
|
*
|
|
1765
|
-
* * Watermark: Monochrome transparent text, logos, images, URLs,
|
|
1854
|
+
* * Watermark: Monochrome transparent text, logos, images, URLs, store names, etc.
|
|
1766
1855
|
*
|
|
1767
|
-
* * Text-containing color block: A color block containing text outside the subject.
|
|
1856
|
+
* * Text-containing color block: A color block containing text outside the main subject.
|
|
1768
1857
|
*
|
|
1769
|
-
* **2. The
|
|
1858
|
+
* **2. The intelligent element recognition API provides the following features:**
|
|
1770
1859
|
* * Element recognition: Identifies whether elements such as text, logos, watermarks, and text-containing color blocks exist in the image subject and background.
|
|
1771
1860
|
*
|
|
1772
1861
|
* * Workflow collaboration: Used in combination with the intelligent elimination API. The intelligent element recognition API first identifies the elements to be removed, and then the intelligent elimination API effectively removes them to ensure the accuracy and efficiency of image processing.
|
|
@@ -1872,15 +1961,15 @@ class Client extends openapi_core_1.default {
|
|
|
1872
1961
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ImageRecognitionAsyncResponse({}));
|
|
1873
1962
|
}
|
|
1874
1963
|
/**
|
|
1875
|
-
* Identifies specific elements in e-commerce images.
|
|
1964
|
+
* Identifies specific elements in e-commerce images. This API extracts image details to help users quickly identify text, logos, watermarks, and text-containing color blocks in the main subject and background, significantly improving the efficiency and accuracy of image filtering.
|
|
1876
1965
|
*
|
|
1877
1966
|
* @remarks
|
|
1878
1967
|
* # Scenarios
|
|
1879
1968
|
* 1. Intelligent image filtering and optimization:
|
|
1880
|
-
* Merchants and platforms can use the intelligent element recognition API to filter images and automatically identify images containing specific elements such as watermarks and logos. After filtering, they can
|
|
1969
|
+
* Merchants and platforms can use the intelligent element recognition API to filter images and automatically identify images containing specific elements such as watermarks and logos. After filtering, they can call the intelligent elimination API to optimize these images, improve image quality, and shorten the product release cycle.
|
|
1881
1970
|
* 2. Automated image processing and efficiency improvement:
|
|
1882
1971
|
* Through the intelligent element recognition API, merchants and platforms can automate the image processing workflow and automatically identify and mark image elements that need to be optimized. This workflow reduces manual intervention and improves the efficiency of image processing.
|
|
1883
|
-
* #
|
|
1972
|
+
* # Feature description
|
|
1884
1973
|
* **1. Definitions of image subject, image background, text, logo, watermark, and text-containing color block: For details, see the relevant examples in the elimination documentation (reserved text link).**
|
|
1885
1974
|
* * Image subject: The core product area in the image.
|
|
1886
1975
|
*
|
|
@@ -1890,11 +1979,11 @@ class Client extends openapi_core_1.default {
|
|
|
1890
1979
|
*
|
|
1891
1980
|
* * Logo: A logo or trademark.
|
|
1892
1981
|
*
|
|
1893
|
-
* * Watermark: Monochrome transparent text, logos, images, URLs,
|
|
1982
|
+
* * Watermark: Monochrome transparent text, logos, images, URLs, store names, etc.
|
|
1894
1983
|
*
|
|
1895
|
-
* * Text-containing color block: A color block containing text outside the subject.
|
|
1984
|
+
* * Text-containing color block: A color block containing text outside the main subject.
|
|
1896
1985
|
*
|
|
1897
|
-
* **2. The
|
|
1986
|
+
* **2. The intelligent element recognition API provides the following features:**
|
|
1898
1987
|
* * Element recognition: Identifies whether elements such as text, logos, watermarks, and text-containing color blocks exist in the image subject and background.
|
|
1899
1988
|
*
|
|
1900
1989
|
* * Workflow collaboration: Used in combination with the intelligent elimination API. The intelligent element recognition API first identifies the elements to be removed, and then the intelligent elimination API effectively removes them to ensure the accuracy and efficiency of image processing.
|
|
@@ -2039,27 +2128,27 @@ class Client extends openapi_core_1.default {
|
|
|
2039
2128
|
return await this.imageRemovalProWithOptions(request, runtime);
|
|
2040
2129
|
}
|
|
2041
2130
|
/**
|
|
2042
|
-
*
|
|
2131
|
+
* Provides intelligent image erasure tailored for e-commerce images, automatically identifying and removing text, specific names, transparent text blocks, and unwanted overlays. It allows you to specify objects to remove as needed, enabling efficient and precise processing of large image batches and simplifying image editing workflows.
|
|
2043
2132
|
*
|
|
2044
2133
|
* @remarks
|
|
2045
2134
|
* ## Scenarios
|
|
2046
|
-
* * **Efficient e-commerce
|
|
2135
|
+
* * **Efficient processing of e-commerce images:**
|
|
2047
2136
|
* Removes irrelevant visual elements to provide clearer and more focused product displays, enhancing the customer shopping experience.
|
|
2048
|
-
* * **
|
|
2049
|
-
*
|
|
2137
|
+
* * **Synchronization of product information across multiple platforms:**
|
|
2138
|
+
* When merchants list products on different e-commerce platforms, they can quickly remove specific text and watermarks to comply with the image specifications of different platforms.
|
|
2050
2139
|
* ## Features
|
|
2051
|
-
* 1. First,
|
|
2140
|
+
* 1. First, clarify the definitions of the image subject and the image background. When specifying the objects to remove, you can choose to remove them from the image subject or the image background. Both API and editor integration methods are supported.
|
|
2052
2141
|
*
|
|
2053
|
-
* 1. Image subject:
|
|
2142
|
+
* 1. Image subject: The core product area in the image.
|
|
2054
2143
|
*
|
|
2055
|
-
* 2. Image background:
|
|
2144
|
+
* 2. Image background: The rest of the image excluding the product subject.
|
|
2056
2145
|
*
|
|
2057
|
-
* 2. Second, the objects to be removed
|
|
2058
|
-
* ##
|
|
2146
|
+
* 2. Second, the objects to be removed support the following four types: text, specific names, transparent text blocks, and watermarks. Refer to the following examples to clarify your requirements and perform targeted operations:
|
|
2147
|
+
* ## Before and after
|
|
2059
2148
|
* | **Original image** | **Processed image** |
|
|
2060
2149
|
* | --- | --- |
|
|
2061
|
-
* |  |  |
|
|
2151
|
+
* |  |  |
|
|
2063
2152
|
*
|
|
2064
2153
|
* @param request - ImageRemovalProAsyncRequest
|
|
2065
2154
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -2091,27 +2180,27 @@ class Client extends openapi_core_1.default {
|
|
|
2091
2180
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ImageRemovalProAsyncResponse({}));
|
|
2092
2181
|
}
|
|
2093
2182
|
/**
|
|
2094
|
-
*
|
|
2183
|
+
* Provides intelligent image erasure tailored for e-commerce images, automatically identifying and removing text, specific names, transparent text blocks, and unwanted overlays. It allows you to specify objects to remove as needed, enabling efficient and precise processing of large image batches and simplifying image editing workflows.
|
|
2095
2184
|
*
|
|
2096
2185
|
* @remarks
|
|
2097
2186
|
* ## Scenarios
|
|
2098
|
-
* * **Efficient e-commerce
|
|
2187
|
+
* * **Efficient processing of e-commerce images:**
|
|
2099
2188
|
* Removes irrelevant visual elements to provide clearer and more focused product displays, enhancing the customer shopping experience.
|
|
2100
|
-
* * **
|
|
2101
|
-
*
|
|
2189
|
+
* * **Synchronization of product information across multiple platforms:**
|
|
2190
|
+
* When merchants list products on different e-commerce platforms, they can quickly remove specific text and watermarks to comply with the image specifications of different platforms.
|
|
2102
2191
|
* ## Features
|
|
2103
|
-
* 1. First,
|
|
2192
|
+
* 1. First, clarify the definitions of the image subject and the image background. When specifying the objects to remove, you can choose to remove them from the image subject or the image background. Both API and editor integration methods are supported.
|
|
2104
2193
|
*
|
|
2105
|
-
* 1. Image subject:
|
|
2194
|
+
* 1. Image subject: The core product area in the image.
|
|
2106
2195
|
*
|
|
2107
|
-
* 2. Image background:
|
|
2196
|
+
* 2. Image background: The rest of the image excluding the product subject.
|
|
2108
2197
|
*
|
|
2109
|
-
* 2. Second, the objects to be removed
|
|
2110
|
-
* ##
|
|
2198
|
+
* 2. Second, the objects to be removed support the following four types: text, specific names, transparent text blocks, and watermarks. Refer to the following examples to clarify your requirements and perform targeted operations:
|
|
2199
|
+
* ## Before and after
|
|
2111
2200
|
* | **Original image** | **Processed image** |
|
|
2112
2201
|
* | --- | --- |
|
|
2113
|
-
* |  |  |
|
|
2203
|
+
* |  |  |
|
|
2115
2204
|
*
|
|
2116
2205
|
* @param request - ImageRemovalProAsyncRequest
|
|
2117
2206
|
* @returns ImageRemovalProAsyncResponse
|
|
@@ -2360,35 +2449,35 @@ class Client extends openapi_core_1.default {
|
|
|
2360
2449
|
return await this.imageRemoveWithOptions(request, runtime);
|
|
2361
2450
|
}
|
|
2362
2451
|
/**
|
|
2363
|
-
*
|
|
2452
|
+
* Provides intelligent object removal tailored for e-commerce images. It automatically identifies and removes text, specific names, transparent text blocks, and watermarks. You can also specify objects for removal as needed, enabling efficient and accurate processing of large image batches and simplifying image editing workflows.
|
|
2364
2453
|
*
|
|
2365
2454
|
* @remarks
|
|
2366
2455
|
* ## Scenarios
|
|
2367
2456
|
* * **Efficient processing of e-commerce images:**
|
|
2368
2457
|
* Removes irrelevant visual elements to provide clearer and more focused product displays, enhancing the customer shopping experience.
|
|
2369
2458
|
* * **Multi-platform product information synchronization:**
|
|
2370
|
-
* When merchants
|
|
2459
|
+
* When merchants list products on different e-commerce platforms, they can quickly remove specific text and watermarks to comply with the image specifications of different platforms.
|
|
2371
2460
|
* * **Marketing material preparation:**
|
|
2372
2461
|
* Merchants can easily remove original promotional information and quickly update it with the latest campaign information.
|
|
2373
|
-
* > Note: Merchants or platforms can
|
|
2462
|
+
* > Note: Merchants or platforms can remove impurities from images in batches to optimize image quality. Use this API in combination with the Intelligent Recognition API (coming soon) to first filter images containing text, specific names, transparent text blocks, and watermarks, and then use the Intelligent Erasure API for batch removal. The Intelligent Recognition API focuses on recognition, while the Intelligent Erasure API focuses on erasure. Calling them in sequence maximizes the erasure effect and reduces false removals and false recognitions.
|
|
2374
2463
|
* ## Features
|
|
2375
|
-
* 1. First, clarify the definitions of the image subject and the image background. In actual use, when specifying the objects to remove, you can choose whether the removal scope is the image subject or the image background
|
|
2464
|
+
* 1. First, clarify the definitions of the image subject and the image background. In actual use, when specifying the objects to remove, you can choose whether the removal scope is the image subject or the image background. Both API and editor integration methods are supported to meet different needs.
|
|
2376
2465
|
*
|
|
2377
2466
|
* 1. Image subject: the core product area in the image.
|
|
2378
2467
|
*
|
|
2379
2468
|
* 2. Image background: the rest of the image other than the product subject.
|
|
2380
2469
|
*
|
|
2381
|
-
* 2. Second, the
|
|
2470
|
+
* 2. Second, the following three types of objects can be removed:
|
|
2382
2471
|
*
|
|
2383
|
-
* 1.
|
|
2472
|
+
* 1. Select by type: You can select text, specific names, transparent text blocks, and watermarks. Refer to the following examples to clarify your requirements and perform targeted operations.
|
|
2384
2473
|
*
|
|
2385
|
-
* 2. (Upgraded)
|
|
2474
|
+
* 2. (Upgraded) Specify images to remove: You can specify multiple images at a time.
|
|
2386
2475
|
*
|
|
2387
|
-
* 3. (Upgraded)
|
|
2476
|
+
* 3. (Upgraded) Specify text to remove: You can specify multiple text strings at a time.
|
|
2388
2477
|
*
|
|
2389
|
-
* Processing flow:
|
|
2478
|
+
* Processing flow: Pass the original image URL -> The OCR and VLM models detect text or image positions -> Determine the coordinates of the area to be inpainted -> AI inpainting generates a clean image -> Return the result image URL.
|
|
2390
2479
|
* ## Effect comparison
|
|
2391
|
-
* | **Objects that can be removed
|
|
2480
|
+
* | **Objects that can be removed** | **Original image** | **Processed image** |
|
|
2392
2481
|
* | --- | --- | --- |
|
|
2393
2482
|
* | Text | * Involves the image subject area
|
|
2394
2483
|
*
|
|
@@ -2438,8 +2527,8 @@ class Client extends openapi_core_1.default {
|
|
|
2438
2527
|
* * Does not involve the image subject area
|
|
2439
2528
|
*
|
|
2440
2529
|
*  |
|
|
2441
|
-
* | Specified image element
|
|
2442
|
-
* | Specified text
|
|
2530
|
+
* | Specified image element removal (supports multiple image URLs) |  |  |
|
|
2531
|
+
* | Specified text removal (supports multiple text strings) |  |  |
|
|
2443
2532
|
*
|
|
2444
2533
|
* @param tmpReq - ImageRemoveAsyncRequest
|
|
2445
2534
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -2503,35 +2592,35 @@ class Client extends openapi_core_1.default {
|
|
|
2503
2592
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ImageRemoveAsyncResponse({}));
|
|
2504
2593
|
}
|
|
2505
2594
|
/**
|
|
2506
|
-
*
|
|
2595
|
+
* Provides intelligent object removal tailored for e-commerce images. It automatically identifies and removes text, specific names, transparent text blocks, and watermarks. You can also specify objects for removal as needed, enabling efficient and accurate processing of large image batches and simplifying image editing workflows.
|
|
2507
2596
|
*
|
|
2508
2597
|
* @remarks
|
|
2509
2598
|
* ## Scenarios
|
|
2510
2599
|
* * **Efficient processing of e-commerce images:**
|
|
2511
2600
|
* Removes irrelevant visual elements to provide clearer and more focused product displays, enhancing the customer shopping experience.
|
|
2512
2601
|
* * **Multi-platform product information synchronization:**
|
|
2513
|
-
* When merchants
|
|
2602
|
+
* When merchants list products on different e-commerce platforms, they can quickly remove specific text and watermarks to comply with the image specifications of different platforms.
|
|
2514
2603
|
* * **Marketing material preparation:**
|
|
2515
2604
|
* Merchants can easily remove original promotional information and quickly update it with the latest campaign information.
|
|
2516
|
-
* > Note: Merchants or platforms can
|
|
2605
|
+
* > Note: Merchants or platforms can remove impurities from images in batches to optimize image quality. Use this API in combination with the Intelligent Recognition API (coming soon) to first filter images containing text, specific names, transparent text blocks, and watermarks, and then use the Intelligent Erasure API for batch removal. The Intelligent Recognition API focuses on recognition, while the Intelligent Erasure API focuses on erasure. Calling them in sequence maximizes the erasure effect and reduces false removals and false recognitions.
|
|
2517
2606
|
* ## Features
|
|
2518
|
-
* 1. First, clarify the definitions of the image subject and the image background. In actual use, when specifying the objects to remove, you can choose whether the removal scope is the image subject or the image background
|
|
2607
|
+
* 1. First, clarify the definitions of the image subject and the image background. In actual use, when specifying the objects to remove, you can choose whether the removal scope is the image subject or the image background. Both API and editor integration methods are supported to meet different needs.
|
|
2519
2608
|
*
|
|
2520
2609
|
* 1. Image subject: the core product area in the image.
|
|
2521
2610
|
*
|
|
2522
2611
|
* 2. Image background: the rest of the image other than the product subject.
|
|
2523
2612
|
*
|
|
2524
|
-
* 2. Second, the
|
|
2613
|
+
* 2. Second, the following three types of objects can be removed:
|
|
2525
2614
|
*
|
|
2526
|
-
* 1.
|
|
2615
|
+
* 1. Select by type: You can select text, specific names, transparent text blocks, and watermarks. Refer to the following examples to clarify your requirements and perform targeted operations.
|
|
2527
2616
|
*
|
|
2528
|
-
* 2. (Upgraded)
|
|
2617
|
+
* 2. (Upgraded) Specify images to remove: You can specify multiple images at a time.
|
|
2529
2618
|
*
|
|
2530
|
-
* 3. (Upgraded)
|
|
2619
|
+
* 3. (Upgraded) Specify text to remove: You can specify multiple text strings at a time.
|
|
2531
2620
|
*
|
|
2532
|
-
* Processing flow:
|
|
2621
|
+
* Processing flow: Pass the original image URL -> The OCR and VLM models detect text or image positions -> Determine the coordinates of the area to be inpainted -> AI inpainting generates a clean image -> Return the result image URL.
|
|
2533
2622
|
* ## Effect comparison
|
|
2534
|
-
* | **Objects that can be removed
|
|
2623
|
+
* | **Objects that can be removed** | **Original image** | **Processed image** |
|
|
2535
2624
|
* | --- | --- | --- |
|
|
2536
2625
|
* | Text | * Involves the image subject area
|
|
2537
2626
|
*
|
|
@@ -2581,8 +2670,8 @@ class Client extends openapi_core_1.default {
|
|
|
2581
2670
|
* * Does not involve the image subject area
|
|
2582
2671
|
*
|
|
2583
2672
|
*  |
|
|
2584
|
-
* | Specified image element
|
|
2585
|
-
* | Specified text
|
|
2673
|
+
* | Specified image element removal (supports multiple image URLs) |  |  |
|
|
2674
|
+
* | Specified text removal (supports multiple text strings) |  |  |
|
|
2586
2675
|
*
|
|
2587
2676
|
* @param request - ImageRemoveAsyncRequest
|
|
2588
2677
|
* @returns ImageRemoveAsyncResponse
|
|
@@ -2705,17 +2794,17 @@ class Client extends openapi_core_1.default {
|
|
|
2705
2794
|
return await this.imageTranslationPlusWithOptions(request, runtime);
|
|
2706
2795
|
}
|
|
2707
2796
|
/**
|
|
2708
|
-
*
|
|
2797
|
+
* Translates e-commerce images by using the Image Translation Plus edition. This edition adopts the Mixture of Experts (MoE) architecture and delivers higher translation accuracy for multiple minor languages compared with the Lite and Pro editions.
|
|
2709
2798
|
*
|
|
2710
2799
|
* @remarks
|
|
2711
2800
|
* ## Scenarios
|
|
2712
2801
|
* Main images and detail images for cross-border e-commerce.
|
|
2713
2802
|
* ## Features
|
|
2714
|
-
* * **Product subject information protection**:
|
|
2803
|
+
* * **Product subject information protection**: Allows you to specify whether to translate the text on the product subject. This helps you protect the subject information, such as the embedded product name, from being translated.
|
|
2715
2804
|
*
|
|
2716
|
-
* * **Brand name protection**:
|
|
2805
|
+
* * **Brand name protection**: Allows you to specify whether to translate the brand names in the image. This helps you protect the brand name information from being translated.
|
|
2717
2806
|
*
|
|
2718
|
-
* * **Translation intervention**:
|
|
2807
|
+
* * **Translation intervention**: Allows you to customize translation results, including keeping the original text (ABC-ABC), specifying a translation (ABC-DEF), and skipping translation (ABC-null). This is commonly used in scenarios such as brand word protection. You can implement this by passing the corresponding intervention glossary ID when you call the API operation. This meets your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention words. If you need more, contact the platform for assistance.
|
|
2719
2808
|
* ## Supported language pairs
|
|
2720
2809
|
* | **No.** | **Source language** | | **Target language** | |
|
|
2721
2810
|
* | --- | --- | --- | --- | --- |
|
|
@@ -2733,8 +2822,8 @@ class Client extends openapi_core_1.default {
|
|
|
2733
2822
|
* | 11 | en | English | de | German |
|
|
2734
2823
|
* | 12 | en | English | pl | Polish |
|
|
2735
2824
|
* ## Preparations and call process
|
|
2736
|
-
* Image Translation Plus and other
|
|
2737
|
-
* For the
|
|
2825
|
+
* The asynchronous API operations for Image Translation Plus and other features share the same result retrieval process, RocketMQ subscription rules, message encryption protocols, dependency configurations, and Java consumption and decryption code.
|
|
2826
|
+
* For the asynchronous call process, subscription rules, message formats and encryption protocols, and detailed integration solutions, refer to [Asynchronous API operation call instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
2738
2827
|
*
|
|
2739
2828
|
* @param request - ImageTranslationPlusAsyncRequest
|
|
2740
2829
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -2784,17 +2873,17 @@ class Client extends openapi_core_1.default {
|
|
|
2784
2873
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ImageTranslationPlusAsyncResponse({}));
|
|
2785
2874
|
}
|
|
2786
2875
|
/**
|
|
2787
|
-
*
|
|
2876
|
+
* Translates e-commerce images by using the Image Translation Plus edition. This edition adopts the Mixture of Experts (MoE) architecture and delivers higher translation accuracy for multiple minor languages compared with the Lite and Pro editions.
|
|
2788
2877
|
*
|
|
2789
2878
|
* @remarks
|
|
2790
2879
|
* ## Scenarios
|
|
2791
2880
|
* Main images and detail images for cross-border e-commerce.
|
|
2792
2881
|
* ## Features
|
|
2793
|
-
* * **Product subject information protection**:
|
|
2882
|
+
* * **Product subject information protection**: Allows you to specify whether to translate the text on the product subject. This helps you protect the subject information, such as the embedded product name, from being translated.
|
|
2794
2883
|
*
|
|
2795
|
-
* * **Brand name protection**:
|
|
2884
|
+
* * **Brand name protection**: Allows you to specify whether to translate the brand names in the image. This helps you protect the brand name information from being translated.
|
|
2796
2885
|
*
|
|
2797
|
-
* * **Translation intervention**:
|
|
2886
|
+
* * **Translation intervention**: Allows you to customize translation results, including keeping the original text (ABC-ABC), specifying a translation (ABC-DEF), and skipping translation (ABC-null). This is commonly used in scenarios such as brand word protection. You can implement this by passing the corresponding intervention glossary ID when you call the API operation. This meets your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention words. If you need more, contact the platform for assistance.
|
|
2798
2887
|
* ## Supported language pairs
|
|
2799
2888
|
* | **No.** | **Source language** | | **Target language** | |
|
|
2800
2889
|
* | --- | --- | --- | --- | --- |
|
|
@@ -2812,8 +2901,8 @@ class Client extends openapi_core_1.default {
|
|
|
2812
2901
|
* | 11 | en | English | de | German |
|
|
2813
2902
|
* | 12 | en | English | pl | Polish |
|
|
2814
2903
|
* ## Preparations and call process
|
|
2815
|
-
* Image Translation Plus and other
|
|
2816
|
-
* For the
|
|
2904
|
+
* The asynchronous API operations for Image Translation Plus and other features share the same result retrieval process, RocketMQ subscription rules, message encryption protocols, dependency configurations, and Java consumption and decryption code.
|
|
2905
|
+
* For the asynchronous call process, subscription rules, message formats and encryption protocols, and detailed integration solutions, refer to [Asynchronous API operation call instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
2817
2906
|
*
|
|
2818
2907
|
* @param request - ImageTranslationPlusAsyncRequest
|
|
2819
2908
|
* @returns ImageTranslationPlusAsyncResponse
|
|
@@ -2907,22 +2996,22 @@ class Client extends openapi_core_1.default {
|
|
|
2907
2996
|
return await this.imageTranslationProWithOptions(request, runtime);
|
|
2908
2997
|
}
|
|
2909
2998
|
/**
|
|
2910
|
-
* Translates images by using Image Translation Pro, which is designed
|
|
2999
|
+
* Translates images by using the Image Translation Pro edition, which is designed for e-commerce images, integrates multimodal large model technology for more accurate image understanding, significantly improves translation quality, and continuously expands and optimizes multilingual translation capabilities, supporting more than 400 language pairs including bridged languages.
|
|
2911
3000
|
*
|
|
2912
3001
|
* @remarks
|
|
2913
3002
|
* ## Scenarios
|
|
2914
|
-
* E-commerce product images, marketing images, and images
|
|
3003
|
+
* E-commerce product images, marketing images, and images in various other scenarios.
|
|
2915
3004
|
* ## Features
|
|
2916
|
-
* * Product subject information protection: You can specify whether to translate the text on the product subject. This helps you protect the subject information, such as the product name, from being translated.
|
|
3005
|
+
* * Product subject information protection: You can specify whether to translate the text on the product subject. This helps you protect the subject information, such as the embedded product name, from being translated.
|
|
2917
3006
|
*
|
|
2918
3007
|
* * Secondary editing after translation: You can specify whether to return layout information such as the position, font, and color of the text. This information can be used for secondary editing when you integrate an image editor. The SDK package for the editor is not yet available. Follow platform notifications for updates.
|
|
2919
3008
|
*
|
|
2920
3009
|
* * Brand name protection: You can specify whether to translate brand names in the image. This helps you protect brand name information from being translated.
|
|
2921
3010
|
*
|
|
2922
|
-
* * Translation intervention: You can customize translation results, including keeping the original text (ABC-ABC), specifying a translation (ABC-DEF), or omitting the translation (ABC-null
|
|
3011
|
+
* * Translation intervention: You can customize translation results, including keeping the original text (ABC-ABC), specifying a translation (ABC-DEF), or omitting the translation (ABC-null). This is commonly used in scenarios such as brand word protection. You can implement this by passing the corresponding intervention glossary ID when calling the API, which meets your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention entries. If you need more, contact the platform for assistance.
|
|
2923
3012
|
* ## Preparations and call process
|
|
2924
3013
|
* Image Translation Pro supports both synchronous and asynchronous call methods. This topic describes the asynchronous call API. For synchronous calls, refer to [Image Translation Pro](https://www.alibabacloud.com/help/en/aidge/image-translation-pro).
|
|
2925
|
-
*
|
|
3014
|
+
* Asynchronous APIs for text translation and other capabilities share the same result retrieval process, RocketMQ subscription rules, message encryption protocols, dependency configurations, and Java consumption and decryption code.
|
|
2926
3015
|
* For the integration process, subscription rules, message formats and encryption protocols, and detailed integration solutions for asynchronous calls, refer to [Asynchronous API call instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
2927
3016
|
*
|
|
2928
3017
|
* @param request - ImageTranslationProAsyncRequest
|
|
@@ -2973,22 +3062,22 @@ class Client extends openapi_core_1.default {
|
|
|
2973
3062
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ImageTranslationProAsyncResponse({}));
|
|
2974
3063
|
}
|
|
2975
3064
|
/**
|
|
2976
|
-
* Translates images by using Image Translation Pro, which is designed
|
|
3065
|
+
* Translates images by using the Image Translation Pro edition, which is designed for e-commerce images, integrates multimodal large model technology for more accurate image understanding, significantly improves translation quality, and continuously expands and optimizes multilingual translation capabilities, supporting more than 400 language pairs including bridged languages.
|
|
2977
3066
|
*
|
|
2978
3067
|
* @remarks
|
|
2979
3068
|
* ## Scenarios
|
|
2980
|
-
* E-commerce product images, marketing images, and images
|
|
3069
|
+
* E-commerce product images, marketing images, and images in various other scenarios.
|
|
2981
3070
|
* ## Features
|
|
2982
|
-
* * Product subject information protection: You can specify whether to translate the text on the product subject. This helps you protect the subject information, such as the product name, from being translated.
|
|
3071
|
+
* * Product subject information protection: You can specify whether to translate the text on the product subject. This helps you protect the subject information, such as the embedded product name, from being translated.
|
|
2983
3072
|
*
|
|
2984
3073
|
* * Secondary editing after translation: You can specify whether to return layout information such as the position, font, and color of the text. This information can be used for secondary editing when you integrate an image editor. The SDK package for the editor is not yet available. Follow platform notifications for updates.
|
|
2985
3074
|
*
|
|
2986
3075
|
* * Brand name protection: You can specify whether to translate brand names in the image. This helps you protect brand name information from being translated.
|
|
2987
3076
|
*
|
|
2988
|
-
* * Translation intervention: You can customize translation results, including keeping the original text (ABC-ABC), specifying a translation (ABC-DEF), or omitting the translation (ABC-null
|
|
3077
|
+
* * Translation intervention: You can customize translation results, including keeping the original text (ABC-ABC), specifying a translation (ABC-DEF), or omitting the translation (ABC-null). This is commonly used in scenarios such as brand word protection. You can implement this by passing the corresponding intervention glossary ID when calling the API, which meets your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention entries. If you need more, contact the platform for assistance.
|
|
2989
3078
|
* ## Preparations and call process
|
|
2990
3079
|
* Image Translation Pro supports both synchronous and asynchronous call methods. This topic describes the asynchronous call API. For synchronous calls, refer to [Image Translation Pro](https://www.alibabacloud.com/help/en/aidge/image-translation-pro).
|
|
2991
|
-
*
|
|
3080
|
+
* Asynchronous APIs for text translation and other capabilities share the same result retrieval process, RocketMQ subscription rules, message encryption protocols, dependency configurations, and Java consumption and decryption code.
|
|
2992
3081
|
* For the integration process, subscription rules, message formats and encryption protocols, and detailed integration solutions for asynchronous calls, refer to [Asynchronous API call instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
2993
3082
|
*
|
|
2994
3083
|
* @param request - ImageTranslationProAsyncRequest
|
|
@@ -2999,22 +3088,22 @@ class Client extends openapi_core_1.default {
|
|
|
2999
3088
|
return await this.imageTranslationProAsyncWithOptions(request, runtime);
|
|
3000
3089
|
}
|
|
3001
3090
|
/**
|
|
3002
|
-
*
|
|
3091
|
+
* Provides image translation tailored for e-commerce, supporting over 400 language pairs, including bridged pairs. Accurately processes complex text layouts within images and ensures translated content aligns with the image context, helping e-commerce platforms and developers easily achieve multilingual conversion and presentation. Additionally, returns layer information such as text position, font, and color to facilitate secondary editing through image editors.
|
|
3003
3092
|
*
|
|
3004
3093
|
* @remarks
|
|
3005
|
-
* # 1. Product
|
|
3006
|
-
* The image translation product is tailored for e-commerce images, supporting over 400 language pairs
|
|
3094
|
+
* # 1. Product introduction
|
|
3095
|
+
* The image translation product is tailored for e-commerce images, supporting over 400 language pairs, including bridged pairs. It accurately processes complex text layouts within images and ensures that the translated content aligns with the image context, helping e-commerce platforms and developers easily achieve multilingual conversion and presentation. Additionally, it returns layer information such as text position, font, and color to facilitate secondary editing through image editors.
|
|
3007
3096
|
* # 2. Scenarios
|
|
3008
3097
|
* E-commerce product images, marketing images, and images for various other scenarios.
|
|
3009
3098
|
* # 3. Features
|
|
3010
|
-
* * **Product subject information protection**: Allows you to
|
|
3099
|
+
* * **Product subject information protection**: Allows you to choose whether to translate the text on the product subject. This helps protect the subject information from being translated, avoiding the translation of embedded information such as product names.
|
|
3011
3100
|
*
|
|
3012
|
-
* * **Brand name protection**: Allows you to
|
|
3101
|
+
* * **Brand name protection**: Allows you to choose whether to translate brand names in the image. This helps protect brand name information from being translated.
|
|
3013
3102
|
*
|
|
3014
|
-
* * **Secondary editing after translation (layer information)**: Allows you to
|
|
3103
|
+
* * **Secondary editing after translation (layer information)**: Allows you to choose whether to return layer information such as text position, font, and color. This information can be used to integrate with image editors for secondary editing of the translation results. For updates on the release of the editor SDK package, refer to platform notifications.
|
|
3015
3104
|
*
|
|
3016
|
-
* * **Translation intervention**: Allows you to customize translation results, including
|
|
3017
|
-
* <props="china">* **Image editor integration**: For more information, see [Image
|
|
3105
|
+
* * **Translation intervention**: Allows you to customize translation results, including do not translate (ABC-ABC), specify translation (ABC-DEF), and skip translation (ABC-null value). This is commonly used in scenarios such as brand word protection. You can achieve this by passing the corresponding intervention glossary ID when calling the API, meeting your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention terms. If you need more, contact the platform for assistance.
|
|
3106
|
+
* <props="china">* **Image editor integration**: For more information, see [Image translation editor protocol](https://www.alibabacloud.com/help/en/document_detail/3042647.html).
|
|
3018
3107
|
*
|
|
3019
3108
|
* @param request - ImageTranslationStandardRequest
|
|
3020
3109
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -3061,22 +3150,22 @@ class Client extends openapi_core_1.default {
|
|
|
3061
3150
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ImageTranslationStandardResponse({}));
|
|
3062
3151
|
}
|
|
3063
3152
|
/**
|
|
3064
|
-
*
|
|
3153
|
+
* Provides image translation tailored for e-commerce, supporting over 400 language pairs, including bridged pairs. Accurately processes complex text layouts within images and ensures translated content aligns with the image context, helping e-commerce platforms and developers easily achieve multilingual conversion and presentation. Additionally, returns layer information such as text position, font, and color to facilitate secondary editing through image editors.
|
|
3065
3154
|
*
|
|
3066
3155
|
* @remarks
|
|
3067
|
-
* # 1. Product
|
|
3068
|
-
* The image translation product is tailored for e-commerce images, supporting over 400 language pairs
|
|
3156
|
+
* # 1. Product introduction
|
|
3157
|
+
* The image translation product is tailored for e-commerce images, supporting over 400 language pairs, including bridged pairs. It accurately processes complex text layouts within images and ensures that the translated content aligns with the image context, helping e-commerce platforms and developers easily achieve multilingual conversion and presentation. Additionally, it returns layer information such as text position, font, and color to facilitate secondary editing through image editors.
|
|
3069
3158
|
* # 2. Scenarios
|
|
3070
3159
|
* E-commerce product images, marketing images, and images for various other scenarios.
|
|
3071
3160
|
* # 3. Features
|
|
3072
|
-
* * **Product subject information protection**: Allows you to
|
|
3161
|
+
* * **Product subject information protection**: Allows you to choose whether to translate the text on the product subject. This helps protect the subject information from being translated, avoiding the translation of embedded information such as product names.
|
|
3073
3162
|
*
|
|
3074
|
-
* * **Brand name protection**: Allows you to
|
|
3163
|
+
* * **Brand name protection**: Allows you to choose whether to translate brand names in the image. This helps protect brand name information from being translated.
|
|
3075
3164
|
*
|
|
3076
|
-
* * **Secondary editing after translation (layer information)**: Allows you to
|
|
3165
|
+
* * **Secondary editing after translation (layer information)**: Allows you to choose whether to return layer information such as text position, font, and color. This information can be used to integrate with image editors for secondary editing of the translation results. For updates on the release of the editor SDK package, refer to platform notifications.
|
|
3077
3166
|
*
|
|
3078
|
-
* * **Translation intervention**: Allows you to customize translation results, including
|
|
3079
|
-
* <props="china">* **Image editor integration**: For more information, see [Image
|
|
3167
|
+
* * **Translation intervention**: Allows you to customize translation results, including do not translate (ABC-ABC), specify translation (ABC-DEF), and skip translation (ABC-null value). This is commonly used in scenarios such as brand word protection. You can achieve this by passing the corresponding intervention glossary ID when calling the API, meeting your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention terms. If you need more, contact the platform for assistance.
|
|
3168
|
+
* <props="china">* **Image editor integration**: For more information, see [Image translation editor protocol](https://www.alibabacloud.com/help/en/document_detail/3042647.html).
|
|
3080
3169
|
*
|
|
3081
3170
|
* @param request - ImageTranslationStandardRequest
|
|
3082
3171
|
* @returns ImageTranslationStandardResponse
|
|
@@ -3086,23 +3175,23 @@ class Client extends openapi_core_1.default {
|
|
|
3086
3175
|
return await this.imageTranslationStandardWithOptions(request, runtime);
|
|
3087
3176
|
}
|
|
3088
3177
|
/**
|
|
3089
|
-
* Translates
|
|
3178
|
+
* Translates e-commerce images and supports over 400 language pairs, including bridged pairs. Accurately processes complex text layouts within images and ensures that the translated content matches the image context, which helps e-commerce platforms and developers easily achieve multilingual conversion and presentation of image content. Additionally, the service returns layer information such as the position, font, and color of text in the image to facilitate secondary editing of translation results by using image editors.
|
|
3090
3179
|
*
|
|
3091
3180
|
* @remarks
|
|
3092
3181
|
* ## Scenarios
|
|
3093
|
-
*
|
|
3182
|
+
* Product images, marketing images, and images for various other scenarios in e-commerce.
|
|
3094
3183
|
* ## Features
|
|
3095
|
-
* * **Product subject information protection**: Allows you to specify whether to translate text on the product subject. This
|
|
3184
|
+
* * **Product subject information protection**: Allows you to specify whether to translate the text on the product subject. This helps protect the subject information from being translated and prevents embedded information such as product names from being translated.
|
|
3096
3185
|
*
|
|
3097
|
-
* * **Brand name protection**: Allows you to specify whether to translate brand names in the image. This
|
|
3186
|
+
* * **Brand name protection**: Allows you to specify whether to translate brand names in the image. This helps protect brand name information from being translated.
|
|
3098
3187
|
*
|
|
3099
|
-
* * **Secondary editing after translation (layer information)**: Allows you to specify whether to return layer information such as
|
|
3188
|
+
* * **Secondary editing after translation (layer information)**: Allows you to specify whether to return layer information such as the position, font, and color of the text. This information can be used to integrate with image editors for secondary editing of translation results. For the release progress of the editor SDK package, refer to platform notifications.
|
|
3100
3189
|
*
|
|
3101
|
-
* * **Translation intervention**: Allows you to customize translation results, including
|
|
3190
|
+
* * **Translation intervention**: Allows you to customize translation results, including no translation (ABC-ABC), specified translation (ABC-DEF), and skip translation (ABC-null value). This is commonly used in scenarios such as brand word protection. You can implement this by passing the corresponding intervention glossary ID when calling the API operation to meet your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention words. If you need more, contact the platform for assistance.
|
|
3102
3191
|
* ## Preparations and call process
|
|
3103
|
-
* * Image Translation Standard Edition supports both synchronous and asynchronous
|
|
3104
|
-
* * The asynchronous
|
|
3105
|
-
* * For the integration process, subscription rules, message formats and encryption protocols, and detailed integration solutions for asynchronous calls, refer to [Asynchronous API call instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
3192
|
+
* * Image Translation Standard Edition supports both synchronous and asynchronous call methods. This topic describes the asynchronous call API operation. For synchronous calls, refer to [Image Translation Standard Edition](https://www.alibabacloud.com/help/en/aidge/image-translation).
|
|
3193
|
+
* * The asynchronous API operations for Image Translation Standard Edition and other features share the same result retrieval process, ApsaraMQ for RocketMQ subscription rules, message encryption protocols, dependency configurations, and Java consumption and decryption code.
|
|
3194
|
+
* * For the integration process, subscription rules, message formats and encryption protocols, and detailed integration solutions for asynchronous calls, refer to [Asynchronous API operation call instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
3106
3195
|
*
|
|
3107
3196
|
* @param request - ImageTranslationStandardAsyncRequest
|
|
3108
3197
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -3152,23 +3241,23 @@ class Client extends openapi_core_1.default {
|
|
|
3152
3241
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ImageTranslationStandardAsyncResponse({}));
|
|
3153
3242
|
}
|
|
3154
3243
|
/**
|
|
3155
|
-
* Translates
|
|
3244
|
+
* Translates e-commerce images and supports over 400 language pairs, including bridged pairs. Accurately processes complex text layouts within images and ensures that the translated content matches the image context, which helps e-commerce platforms and developers easily achieve multilingual conversion and presentation of image content. Additionally, the service returns layer information such as the position, font, and color of text in the image to facilitate secondary editing of translation results by using image editors.
|
|
3156
3245
|
*
|
|
3157
3246
|
* @remarks
|
|
3158
3247
|
* ## Scenarios
|
|
3159
|
-
*
|
|
3248
|
+
* Product images, marketing images, and images for various other scenarios in e-commerce.
|
|
3160
3249
|
* ## Features
|
|
3161
|
-
* * **Product subject information protection**: Allows you to specify whether to translate text on the product subject. This
|
|
3250
|
+
* * **Product subject information protection**: Allows you to specify whether to translate the text on the product subject. This helps protect the subject information from being translated and prevents embedded information such as product names from being translated.
|
|
3162
3251
|
*
|
|
3163
|
-
* * **Brand name protection**: Allows you to specify whether to translate brand names in the image. This
|
|
3252
|
+
* * **Brand name protection**: Allows you to specify whether to translate brand names in the image. This helps protect brand name information from being translated.
|
|
3164
3253
|
*
|
|
3165
|
-
* * **Secondary editing after translation (layer information)**: Allows you to specify whether to return layer information such as
|
|
3254
|
+
* * **Secondary editing after translation (layer information)**: Allows you to specify whether to return layer information such as the position, font, and color of the text. This information can be used to integrate with image editors for secondary editing of translation results. For the release progress of the editor SDK package, refer to platform notifications.
|
|
3166
3255
|
*
|
|
3167
|
-
* * **Translation intervention**: Allows you to customize translation results, including
|
|
3256
|
+
* * **Translation intervention**: Allows you to customize translation results, including no translation (ABC-ABC), specified translation (ABC-DEF), and skip translation (ABC-null value). This is commonly used in scenarios such as brand word protection. You can implement this by passing the corresponding intervention glossary ID when calling the API operation to meet your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention words. If you need more, contact the platform for assistance.
|
|
3168
3257
|
* ## Preparations and call process
|
|
3169
|
-
* * Image Translation Standard Edition supports both synchronous and asynchronous
|
|
3170
|
-
* * The asynchronous
|
|
3171
|
-
* * For the integration process, subscription rules, message formats and encryption protocols, and detailed integration solutions for asynchronous calls, refer to [Asynchronous API call instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
3258
|
+
* * Image Translation Standard Edition supports both synchronous and asynchronous call methods. This topic describes the asynchronous call API operation. For synchronous calls, refer to [Image Translation Standard Edition](https://www.alibabacloud.com/help/en/aidge/image-translation).
|
|
3259
|
+
* * The asynchronous API operations for Image Translation Standard Edition and other features share the same result retrieval process, ApsaraMQ for RocketMQ subscription rules, message encryption protocols, dependency configurations, and Java consumption and decryption code.
|
|
3260
|
+
* * For the integration process, subscription rules, message formats and encryption protocols, and detailed integration solutions for asynchronous calls, refer to [Asynchronous API operation call instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
3172
3261
|
*
|
|
3173
3262
|
* @param request - ImageTranslationStandardAsyncRequest
|
|
3174
3263
|
* @returns ImageTranslationStandardAsyncResponse
|
|
@@ -3178,22 +3267,24 @@ class Client extends openapi_core_1.default {
|
|
|
3178
3267
|
return await this.imageTranslationStandardAsyncWithOptions(request, runtime);
|
|
3179
3268
|
}
|
|
3180
3269
|
/**
|
|
3181
|
-
* Upscales e-commerce images by intelligently enhancing resolution and details, improving textures and image quality during the upscaling process. Supports high-definition outputs such as 2K and 4K to meet the requirements for product main images, detail images, and marketing materials, and
|
|
3270
|
+
* Upscales e-commerce images by intelligently enhancing resolution and details, improving textures and image quality during the upscaling process. Supports high-definition outputs such as 2K and 4K to meet the requirements for product main images, detail images, and marketing materials, enabling efficient and clear processing of large batches of images to improve product image quality. (Synchronous)
|
|
3182
3271
|
*
|
|
3183
3272
|
* @remarks
|
|
3184
|
-
* # Product
|
|
3185
|
-
* The high-definition upscaling product is customized for e-commerce images. It intelligently enhances image resolution and
|
|
3273
|
+
* # Product Introduction
|
|
3274
|
+
* The high-definition upscaling product is customized for e-commerce images. It intelligently enhances image resolution and detail performance, and provides texture enhancement and image quality optimization during the upscaling procedure. It supports high-definition outputs such as 2K and 4K to meet the high-definition requirements for product main images, detail images, and marketing materials. It enables efficient and clear processing of large batches of images to improve product image quality. (synchronous)
|
|
3186
3275
|
* # Scenarios
|
|
3187
|
-
* * High-definition e-commerce product images: Enhances the resolution and details of product main images, detail images, and marketing materials to make textures and materials clearer
|
|
3188
|
-
*
|
|
3189
|
-
* *
|
|
3276
|
+
* * High-definition e-commerce product images: Enhances the resolution and details of product main images, detail images, and marketing materials to make details such as product textures and materials clearer, improving the quality of product displays.
|
|
3277
|
+
*
|
|
3278
|
+
* * Multi-platform product material adaptation: Upscales existing product images in high definition to meet the image display requirements of different e-commerce platforms, satisfying the material usage needs across multiple platforms and scenarios.
|
|
3279
|
+
*
|
|
3280
|
+
* * Batch optimization of existing images: Enhances existing low-resolution product images in high definition to reduce reshooting and production costs, improving the reuse efficiency of image materials.
|
|
3190
3281
|
* # Features
|
|
3191
|
-
* 1. Intelligent high-definition upscaling: Based on AI image enhancement technology, it improves the resolution and optimizes the details of input images to enhance image clarity and texture
|
|
3282
|
+
* 1. Intelligent high-definition upscaling: Based on AI image enhancement technology, it improves the resolution and optimizes the details of input images to enhance image clarity and texture performance.
|
|
3192
3283
|
*
|
|
3193
3284
|
* 2. Multiple upscaling factors: Supports 2x and 4x upscaling factors, allowing flexible selection based on image quality and business requirements.
|
|
3194
3285
|
*
|
|
3195
3286
|
* 3. Synchronous API calls: Supports synchronous submission of images and retrieval of processing results through APIs, facilitating integration into e-commerce image processing workflows.
|
|
3196
|
-
* #
|
|
3287
|
+
* # Before and after
|
|
3197
3288
|
* | **Original image** | **Upscaled image** |
|
|
3198
3289
|
* | --- | --- |
|
|
3199
3290
|
* |  |  |
|
|
@@ -3228,22 +3319,24 @@ class Client extends openapi_core_1.default {
|
|
|
3228
3319
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ImageUpscaleResponse({}));
|
|
3229
3320
|
}
|
|
3230
3321
|
/**
|
|
3231
|
-
* Upscales e-commerce images by intelligently enhancing resolution and details, improving textures and image quality during the upscaling process. Supports high-definition outputs such as 2K and 4K to meet the requirements for product main images, detail images, and marketing materials, and
|
|
3322
|
+
* Upscales e-commerce images by intelligently enhancing resolution and details, improving textures and image quality during the upscaling process. Supports high-definition outputs such as 2K and 4K to meet the requirements for product main images, detail images, and marketing materials, enabling efficient and clear processing of large batches of images to improve product image quality. (Synchronous)
|
|
3232
3323
|
*
|
|
3233
3324
|
* @remarks
|
|
3234
|
-
* # Product
|
|
3235
|
-
* The high-definition upscaling product is customized for e-commerce images. It intelligently enhances image resolution and
|
|
3325
|
+
* # Product Introduction
|
|
3326
|
+
* The high-definition upscaling product is customized for e-commerce images. It intelligently enhances image resolution and detail performance, and provides texture enhancement and image quality optimization during the upscaling procedure. It supports high-definition outputs such as 2K and 4K to meet the high-definition requirements for product main images, detail images, and marketing materials. It enables efficient and clear processing of large batches of images to improve product image quality. (synchronous)
|
|
3236
3327
|
* # Scenarios
|
|
3237
|
-
* * High-definition e-commerce product images: Enhances the resolution and details of product main images, detail images, and marketing materials to make textures and materials clearer
|
|
3238
|
-
*
|
|
3239
|
-
* *
|
|
3328
|
+
* * High-definition e-commerce product images: Enhances the resolution and details of product main images, detail images, and marketing materials to make details such as product textures and materials clearer, improving the quality of product displays.
|
|
3329
|
+
*
|
|
3330
|
+
* * Multi-platform product material adaptation: Upscales existing product images in high definition to meet the image display requirements of different e-commerce platforms, satisfying the material usage needs across multiple platforms and scenarios.
|
|
3331
|
+
*
|
|
3332
|
+
* * Batch optimization of existing images: Enhances existing low-resolution product images in high definition to reduce reshooting and production costs, improving the reuse efficiency of image materials.
|
|
3240
3333
|
* # Features
|
|
3241
|
-
* 1. Intelligent high-definition upscaling: Based on AI image enhancement technology, it improves the resolution and optimizes the details of input images to enhance image clarity and texture
|
|
3334
|
+
* 1. Intelligent high-definition upscaling: Based on AI image enhancement technology, it improves the resolution and optimizes the details of input images to enhance image clarity and texture performance.
|
|
3242
3335
|
*
|
|
3243
3336
|
* 2. Multiple upscaling factors: Supports 2x and 4x upscaling factors, allowing flexible selection based on image quality and business requirements.
|
|
3244
3337
|
*
|
|
3245
3338
|
* 3. Synchronous API calls: Supports synchronous submission of images and retrieval of processing results through APIs, facilitating integration into e-commerce image processing workflows.
|
|
3246
|
-
* #
|
|
3339
|
+
* # Before and after
|
|
3247
3340
|
* | **Original image** | **Upscaled image** |
|
|
3248
3341
|
* | --- | --- |
|
|
3249
3342
|
* |  |  |
|
|
@@ -3306,29 +3399,29 @@ class Client extends openapi_core_1.default {
|
|
|
3306
3399
|
return await this.languageDetectWithOptions(request, runtime);
|
|
3307
3400
|
}
|
|
3308
3401
|
/**
|
|
3309
|
-
*
|
|
3402
|
+
* Inspects the compliance of in-store marketing material displays based on a multimodal large language model (MLLM) for offline retail and channel store inspection scenarios. It evaluates standard materials such as functional table cards, posters, and door-shaped display stands, along with long-tail materials like light cloths, KT boards, floor stickers, and light boxes in actual store photos against inspection rules at the natural language level. The operation outputs structured conclusions and Chinese evidence descriptions that can be directly integrated into business systems. This is a **synchronous** call that returns results in a single request.
|
|
3310
3403
|
*
|
|
3311
3404
|
* @remarks
|
|
3312
3405
|
* ## **Scenarios**
|
|
3313
|
-
* - **
|
|
3406
|
+
* - **Store marketing material compliance inspection**: Automatically determines whether specified marketing materials (feature table cards, posters, door-shaped display stands, etc.) are placed or posted in the store according to the standard display guidelines issued by the headquarters. Identifies typical issues such as missing materials, placement faults, and content mismatch.
|
|
3314
3407
|
*
|
|
3315
|
-
* - **New product
|
|
3408
|
+
* - **New product or campaign material verification**: During new product releases or sales promotions, automatically verifies in batches the listing of campaign materials in actual store photos as required, replacing manual sampling inspections.
|
|
3316
3409
|
*
|
|
3317
|
-
* - **Adaptive dual-image comparison and single-image detection**: A single API supports both
|
|
3410
|
+
* - **Adaptive dual-image comparison and single-image detection**: A single API operation supports both dual-image comparison (reference image + target image, Pattern A) and single-image detection (target image only, Pattern B). An automatic switchover is performed based on whether `ImageRefer` is provided, so callers do not need to distinguish between invoke methods.
|
|
3318
3411
|
* ## **Features**
|
|
3319
|
-
* - **Multi-mode intelligent routing**: Built-in material type parsing capability
|
|
3412
|
+
* - **Multi-mode intelligent routing**: Built-in material type parsing capability. Automatically identifies the target material type based on the natural language text in `Rules` and routes the request to the corresponding material detection pipeline. Callers only need to pass the image URL and rule text. Currently supports routing for four detection items: functional table cards, posters, door-shaped display stands, and other materials. More specific marketing material detection pipelines will be added in the future.
|
|
3320
3413
|
*
|
|
3321
|
-
* - **MLLM semantic
|
|
3414
|
+
* - **MLLM semantic understanding + structured rule collaboration**: Uses a multimodal large language model to perform semantic-level understanding such as material recognition, content comparison, and text OCR. Combines rule cleaning and structured models to break down natural language rules into step-by-step verifiable steps (S1, S2, etc.), ensuring accuracy while maintaining rule flexibility and traceability.
|
|
3322
3415
|
*
|
|
3323
|
-
* - **Structured
|
|
3416
|
+
* - **Structured audit conclusion output**: The output is unified into the structure of `Result.OverallResult` + `Result.Steps[]` + `Result.Evidence`. The overall conclusion is derived from the logical AND of all steps. Each step is independently visible, making it easy to directly drive downstream business systems and support case-level manual review tracing.
|
|
3324
3417
|
*
|
|
3325
|
-
* - **Supported input formats**: Currently supports publicly accessible image URLs.
|
|
3326
|
-
* ## **
|
|
3327
|
-
* - **Synchronous call**:
|
|
3418
|
+
* - **Supported input formats**: Currently supports publicly accessible image URLs. Supports both single-image (Mode B) and dual-image (Mode A) calling formats, with completely identical output structures.
|
|
3419
|
+
* ## **Invoke methods**
|
|
3420
|
+
* - **Synchronous call**: Returns the detection result in a single request without the need for polling. The response uses a unified envelope containing `Code`, `Message`, `RequestId`, `Success`, and `Data`.
|
|
3328
3421
|
*
|
|
3329
|
-
* - **Authentication and signing**: Invoked through the Aidge gateway. Authentication, signing, and common parameters follow the platform
|
|
3422
|
+
* - **Authentication and signing**: Invoked through the Aidge gateway. Authentication, signing, and common parameters follow the unified platform access methods. Refer to the officially published API documentation for the specific request path.
|
|
3330
3423
|
*
|
|
3331
|
-
* - **Timeout settings**: Set the request timeout to no less than the maximum response time of the API
|
|
3424
|
+
* - **Timeout settings**: Set the request timeout to a value no less than the maximum response time of the API operation. Refer to the officially published documentation for the specific value.
|
|
3332
3425
|
*
|
|
3333
3426
|
* @param request - MaterialInspectionRequest
|
|
3334
3427
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -3369,29 +3462,29 @@ class Client extends openapi_core_1.default {
|
|
|
3369
3462
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.MaterialInspectionResponse({}));
|
|
3370
3463
|
}
|
|
3371
3464
|
/**
|
|
3372
|
-
*
|
|
3465
|
+
* Inspects the compliance of in-store marketing material displays based on a multimodal large language model (MLLM) for offline retail and channel store inspection scenarios. It evaluates standard materials such as functional table cards, posters, and door-shaped display stands, along with long-tail materials like light cloths, KT boards, floor stickers, and light boxes in actual store photos against inspection rules at the natural language level. The operation outputs structured conclusions and Chinese evidence descriptions that can be directly integrated into business systems. This is a **synchronous** call that returns results in a single request.
|
|
3373
3466
|
*
|
|
3374
3467
|
* @remarks
|
|
3375
3468
|
* ## **Scenarios**
|
|
3376
|
-
* - **
|
|
3469
|
+
* - **Store marketing material compliance inspection**: Automatically determines whether specified marketing materials (feature table cards, posters, door-shaped display stands, etc.) are placed or posted in the store according to the standard display guidelines issued by the headquarters. Identifies typical issues such as missing materials, placement faults, and content mismatch.
|
|
3377
3470
|
*
|
|
3378
|
-
* - **New product
|
|
3471
|
+
* - **New product or campaign material verification**: During new product releases or sales promotions, automatically verifies in batches the listing of campaign materials in actual store photos as required, replacing manual sampling inspections.
|
|
3379
3472
|
*
|
|
3380
|
-
* - **Adaptive dual-image comparison and single-image detection**: A single API supports both
|
|
3473
|
+
* - **Adaptive dual-image comparison and single-image detection**: A single API operation supports both dual-image comparison (reference image + target image, Pattern A) and single-image detection (target image only, Pattern B). An automatic switchover is performed based on whether `ImageRefer` is provided, so callers do not need to distinguish between invoke methods.
|
|
3381
3474
|
* ## **Features**
|
|
3382
|
-
* - **Multi-mode intelligent routing**: Built-in material type parsing capability
|
|
3475
|
+
* - **Multi-mode intelligent routing**: Built-in material type parsing capability. Automatically identifies the target material type based on the natural language text in `Rules` and routes the request to the corresponding material detection pipeline. Callers only need to pass the image URL and rule text. Currently supports routing for four detection items: functional table cards, posters, door-shaped display stands, and other materials. More specific marketing material detection pipelines will be added in the future.
|
|
3383
3476
|
*
|
|
3384
|
-
* - **MLLM semantic
|
|
3477
|
+
* - **MLLM semantic understanding + structured rule collaboration**: Uses a multimodal large language model to perform semantic-level understanding such as material recognition, content comparison, and text OCR. Combines rule cleaning and structured models to break down natural language rules into step-by-step verifiable steps (S1, S2, etc.), ensuring accuracy while maintaining rule flexibility and traceability.
|
|
3385
3478
|
*
|
|
3386
|
-
* - **Structured
|
|
3479
|
+
* - **Structured audit conclusion output**: The output is unified into the structure of `Result.OverallResult` + `Result.Steps[]` + `Result.Evidence`. The overall conclusion is derived from the logical AND of all steps. Each step is independently visible, making it easy to directly drive downstream business systems and support case-level manual review tracing.
|
|
3387
3480
|
*
|
|
3388
|
-
* - **Supported input formats**: Currently supports publicly accessible image URLs.
|
|
3389
|
-
* ## **
|
|
3390
|
-
* - **Synchronous call**:
|
|
3481
|
+
* - **Supported input formats**: Currently supports publicly accessible image URLs. Supports both single-image (Mode B) and dual-image (Mode A) calling formats, with completely identical output structures.
|
|
3482
|
+
* ## **Invoke methods**
|
|
3483
|
+
* - **Synchronous call**: Returns the detection result in a single request without the need for polling. The response uses a unified envelope containing `Code`, `Message`, `RequestId`, `Success`, and `Data`.
|
|
3391
3484
|
*
|
|
3392
|
-
* - **Authentication and signing**: Invoked through the Aidge gateway. Authentication, signing, and common parameters follow the platform
|
|
3485
|
+
* - **Authentication and signing**: Invoked through the Aidge gateway. Authentication, signing, and common parameters follow the unified platform access methods. Refer to the officially published API documentation for the specific request path.
|
|
3393
3486
|
*
|
|
3394
|
-
* - **Timeout settings**: Set the request timeout to no less than the maximum response time of the API
|
|
3487
|
+
* - **Timeout settings**: Set the request timeout to a value no less than the maximum response time of the API operation. Refer to the officially published documentation for the specific value.
|
|
3395
3488
|
*
|
|
3396
3489
|
* @param request - MaterialInspectionRequest
|
|
3397
3490
|
* @returns MaterialInspectionResponse
|
|
@@ -3401,25 +3494,25 @@ class Client extends openapi_core_1.default {
|
|
|
3401
3494
|
return await this.materialInspectionWithOptions(request, runtime);
|
|
3402
3495
|
}
|
|
3403
3496
|
/**
|
|
3404
|
-
* Detects
|
|
3497
|
+
* Detects product displays in physical stores based on multimodal large language models (MLLMs) and visual vector recall. This capability is designed for shelf and floor display inspection scenarios in offline retail and channel stores. The service first uses an MLLM to detect the bounding boxes and preliminary names of each product in real-world shelf and floor display photos, and then performs similarity recall against a customer-specific SKU vector database, enabling an exact match for each product to the customer\\"s own SKUs. It outputs item-by-item structured results that can be directly integrated with business systems. This service is suitable for large-scale, highly concurrent visual recognition pipelines such as remote auditing of offline store displays by headquarters, SKU distribution verification, and display compliance inspections.
|
|
3405
3498
|
*
|
|
3406
3499
|
* @remarks
|
|
3407
3500
|
* ## **Product Introduction**
|
|
3408
|
-
* Display detection is an intelligent
|
|
3501
|
+
* Display detection is an intelligent recognition capability for store product displays built on multimodal large language models (MLLMs) and visual vector recall. It is designed for shelf and floor display inspection scenarios in offline retail and channel stores. For real-world shelf and floor display photos taken in stores, the service first uses an MLLM to detect the bounding boxes and preliminary names of each product, and then performs similarity recall against a customer-specific SKU vector database, enabling an exact match for each product to the customer\\"s own SKUs. It outputs item-by-item structured results that can be directly integrated with business systems. This service is suitable for large-scale, highly concurrent visual recognition pipelines such as remote auditing of offline store displays by headquarters, SKU distribution verification, and display compliance inspections.
|
|
3409
3502
|
* ## **Scenarios**
|
|
3410
|
-
* - **Shelf
|
|
3503
|
+
* - **Shelf and floor display recognition and inventory**: Recognizes each product in a shelf or floor display image and matches it to the customer\\"s SKU database. It outputs an item-by-item list (product name, location, and similarity score) to replace manual item-by-item counting.
|
|
3411
3504
|
*
|
|
3412
|
-
* - **Distribution verification
|
|
3505
|
+
* - **Distribution verification and SKU compliance**: Verifies whether stores display specified SKUs as required by headquarters, and checks for out-of-stock or misplaced items. This supports batch automatic verification during sales promotions or new product launches.
|
|
3413
3506
|
*
|
|
3414
|
-
* - **Customer-specific SKU
|
|
3415
|
-
* ## **
|
|
3416
|
-
* - **Two-stage detection
|
|
3507
|
+
* - **Customer-specific SKU database**: Customers can upload their own standard SKU images to create a dedicated vector database (isolated by RagId). During recognition, retrieval is performed only within this customer\\"s database to ensure accurate naming and data isolation between customers.
|
|
3508
|
+
* ## **Features**
|
|
3509
|
+
* - **Two-stage collaboration of detection and retrieval**: The MLLM detects the bounding box and preliminary name of each product in the entire image (semantic-level understanding, adapted for dense displays and partial occlusion). Then, visual embedding vectors are used to perform similarity retrieval within the customer\\"s SKU database. The preliminary name is replaced with the exact SKU name of the top-1 result, balancing the recall rate and naming accuracy.
|
|
3417
3510
|
*
|
|
3418
|
-
* - **Isolated retrieval with dedicated vector
|
|
3511
|
+
* - **Isolated retrieval with dedicated vector databases**: A dedicated SKU vector database is maintained for each customer based on RagId. Recognition retrieves data only within the corresponding database, ensuring strict isolation between customers. Added or updated SKUs take effect in real time.
|
|
3419
3512
|
*
|
|
3420
|
-
* - **
|
|
3513
|
+
* - **Item-by-item structured output**: The output is unified as Result.Items[]. Each product independently provides the bounding box, retrieved SKU name, SKU code, and similarity score. This facilitates direct integration with downstream inventory and verification systems and supports case-level manual review and tracing.
|
|
3421
3514
|
*
|
|
3422
|
-
* - **Supported input formats**: Currently supports image URLs accessible over the Internet or
|
|
3515
|
+
* - **Supported input formats**: Currently supports image URLs accessible over the Internet or from OSS. A single shelf or floor display image is sufficient for recognition, and the output structure is unified.
|
|
3423
3516
|
*
|
|
3424
3517
|
* @param request - MerchandisePlacementDetectionRequest
|
|
3425
3518
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -3460,25 +3553,25 @@ class Client extends openapi_core_1.default {
|
|
|
3460
3553
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.MerchandisePlacementDetectionResponse({}));
|
|
3461
3554
|
}
|
|
3462
3555
|
/**
|
|
3463
|
-
* Detects
|
|
3556
|
+
* Detects product displays in physical stores based on multimodal large language models (MLLMs) and visual vector recall. This capability is designed for shelf and floor display inspection scenarios in offline retail and channel stores. The service first uses an MLLM to detect the bounding boxes and preliminary names of each product in real-world shelf and floor display photos, and then performs similarity recall against a customer-specific SKU vector database, enabling an exact match for each product to the customer\\"s own SKUs. It outputs item-by-item structured results that can be directly integrated with business systems. This service is suitable for large-scale, highly concurrent visual recognition pipelines such as remote auditing of offline store displays by headquarters, SKU distribution verification, and display compliance inspections.
|
|
3464
3557
|
*
|
|
3465
3558
|
* @remarks
|
|
3466
3559
|
* ## **Product Introduction**
|
|
3467
|
-
* Display detection is an intelligent
|
|
3560
|
+
* Display detection is an intelligent recognition capability for store product displays built on multimodal large language models (MLLMs) and visual vector recall. It is designed for shelf and floor display inspection scenarios in offline retail and channel stores. For real-world shelf and floor display photos taken in stores, the service first uses an MLLM to detect the bounding boxes and preliminary names of each product, and then performs similarity recall against a customer-specific SKU vector database, enabling an exact match for each product to the customer\\"s own SKUs. It outputs item-by-item structured results that can be directly integrated with business systems. This service is suitable for large-scale, highly concurrent visual recognition pipelines such as remote auditing of offline store displays by headquarters, SKU distribution verification, and display compliance inspections.
|
|
3468
3561
|
* ## **Scenarios**
|
|
3469
|
-
* - **Shelf
|
|
3562
|
+
* - **Shelf and floor display recognition and inventory**: Recognizes each product in a shelf or floor display image and matches it to the customer\\"s SKU database. It outputs an item-by-item list (product name, location, and similarity score) to replace manual item-by-item counting.
|
|
3470
3563
|
*
|
|
3471
|
-
* - **Distribution verification
|
|
3564
|
+
* - **Distribution verification and SKU compliance**: Verifies whether stores display specified SKUs as required by headquarters, and checks for out-of-stock or misplaced items. This supports batch automatic verification during sales promotions or new product launches.
|
|
3472
3565
|
*
|
|
3473
|
-
* - **Customer-specific SKU
|
|
3474
|
-
* ## **
|
|
3475
|
-
* - **Two-stage detection
|
|
3566
|
+
* - **Customer-specific SKU database**: Customers can upload their own standard SKU images to create a dedicated vector database (isolated by RagId). During recognition, retrieval is performed only within this customer\\"s database to ensure accurate naming and data isolation between customers.
|
|
3567
|
+
* ## **Features**
|
|
3568
|
+
* - **Two-stage collaboration of detection and retrieval**: The MLLM detects the bounding box and preliminary name of each product in the entire image (semantic-level understanding, adapted for dense displays and partial occlusion). Then, visual embedding vectors are used to perform similarity retrieval within the customer\\"s SKU database. The preliminary name is replaced with the exact SKU name of the top-1 result, balancing the recall rate and naming accuracy.
|
|
3476
3569
|
*
|
|
3477
|
-
* - **Isolated retrieval with dedicated vector
|
|
3570
|
+
* - **Isolated retrieval with dedicated vector databases**: A dedicated SKU vector database is maintained for each customer based on RagId. Recognition retrieves data only within the corresponding database, ensuring strict isolation between customers. Added or updated SKUs take effect in real time.
|
|
3478
3571
|
*
|
|
3479
|
-
* - **
|
|
3572
|
+
* - **Item-by-item structured output**: The output is unified as Result.Items[]. Each product independently provides the bounding box, retrieved SKU name, SKU code, and similarity score. This facilitates direct integration with downstream inventory and verification systems and supports case-level manual review and tracing.
|
|
3480
3573
|
*
|
|
3481
|
-
* - **Supported input formats**: Currently supports image URLs accessible over the Internet or
|
|
3574
|
+
* - **Supported input formats**: Currently supports image URLs accessible over the Internet or from OSS. A single shelf or floor display image is sufficient for recognition, and the output structure is unified.
|
|
3482
3575
|
*
|
|
3483
3576
|
* @param request - MerchandisePlacementDetectionRequest
|
|
3484
3577
|
* @returns MerchandisePlacementDetectionResponse
|
|
@@ -3488,18 +3581,18 @@ class Client extends openapi_core_1.default {
|
|
|
3488
3581
|
return await this.merchandisePlacementDetectionWithOptions(request, runtime);
|
|
3489
3582
|
}
|
|
3490
3583
|
/**
|
|
3491
|
-
* Detects product displays in photos of shelves, endcaps, freezers, display
|
|
3584
|
+
* Detects product displays in photos of shelves, endcaps, freezers, display stands, and dump bins by using product SKU recognition and multimodal large model grounding, aggregates product locations by SKU, and supplements the results with visible single-item grounding and short names to facilitate product location display, inventory counting, and distribution verification.
|
|
3492
3585
|
*
|
|
3493
3586
|
* @remarks
|
|
3494
3587
|
* ## Scenarios
|
|
3495
|
-
* - Display
|
|
3496
|
-
* - Distribution
|
|
3588
|
+
* - Display inventory: Aggregates recognized product bounding boxes by SKU for business system statistics.
|
|
3589
|
+
* - Distribution checklist verification: Compares the recognized SKU list with the target checklist on the business side.
|
|
3497
3590
|
* - Single-item location review: Assists in checking recognition results by combining product bounding boxes and visible names.
|
|
3498
3591
|
* ## Features
|
|
3499
|
-
* - Aggregation by SKU: Aggregates multiple
|
|
3500
|
-
* - Supplementary single-item grounding: The multimodal model
|
|
3592
|
+
* - Aggregation by SKU: Aggregates multiple bounding boxes for the same SkuId in Products, and returns a deduplicated list of SKU codes in skuIds.
|
|
3593
|
+
* - Supplementary single-item grounding: The multimodal model locates each recognizable sales package and provides a confirmable short name.
|
|
3501
3594
|
* - Independent use of two result sets: Products comes from SKU recognition, and grounding comes from supplementary grounding. Grounding does not return SKU codes and does not guarantee item-by-item correspondence with Products.
|
|
3502
|
-
* - Supports single HTTPS image input. Supplementary grounding returns up to 12 targets
|
|
3595
|
+
* - Supports a single HTTPS image as input. Supplementary grounding returns up to 12 targets. Each target includes a category, a bounding box, and visible text of up to 24 characters. The coverage field indicates whether the grounding is complete.
|
|
3503
3596
|
*
|
|
3504
3597
|
* @param request - MerchandisePlacementDetectionLiteRequest
|
|
3505
3598
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -3528,18 +3621,18 @@ class Client extends openapi_core_1.default {
|
|
|
3528
3621
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.MerchandisePlacementDetectionLiteResponse({}));
|
|
3529
3622
|
}
|
|
3530
3623
|
/**
|
|
3531
|
-
* Detects product displays in photos of shelves, endcaps, freezers, display
|
|
3624
|
+
* Detects product displays in photos of shelves, endcaps, freezers, display stands, and dump bins by using product SKU recognition and multimodal large model grounding, aggregates product locations by SKU, and supplements the results with visible single-item grounding and short names to facilitate product location display, inventory counting, and distribution verification.
|
|
3532
3625
|
*
|
|
3533
3626
|
* @remarks
|
|
3534
3627
|
* ## Scenarios
|
|
3535
|
-
* - Display
|
|
3536
|
-
* - Distribution
|
|
3628
|
+
* - Display inventory: Aggregates recognized product bounding boxes by SKU for business system statistics.
|
|
3629
|
+
* - Distribution checklist verification: Compares the recognized SKU list with the target checklist on the business side.
|
|
3537
3630
|
* - Single-item location review: Assists in checking recognition results by combining product bounding boxes and visible names.
|
|
3538
3631
|
* ## Features
|
|
3539
|
-
* - Aggregation by SKU: Aggregates multiple
|
|
3540
|
-
* - Supplementary single-item grounding: The multimodal model
|
|
3632
|
+
* - Aggregation by SKU: Aggregates multiple bounding boxes for the same SkuId in Products, and returns a deduplicated list of SKU codes in skuIds.
|
|
3633
|
+
* - Supplementary single-item grounding: The multimodal model locates each recognizable sales package and provides a confirmable short name.
|
|
3541
3634
|
* - Independent use of two result sets: Products comes from SKU recognition, and grounding comes from supplementary grounding. Grounding does not return SKU codes and does not guarantee item-by-item correspondence with Products.
|
|
3542
|
-
* - Supports single HTTPS image input. Supplementary grounding returns up to 12 targets
|
|
3635
|
+
* - Supports a single HTTPS image as input. Supplementary grounding returns up to 12 targets. Each target includes a category, a bounding box, and visible text of up to 24 characters. The coverage field indicates whether the grounding is complete.
|
|
3543
3636
|
*
|
|
3544
3637
|
* @param request - MerchandisePlacementDetectionLiteRequest
|
|
3545
3638
|
* @returns MerchandisePlacementDetectionLiteResponse
|
|
@@ -3549,18 +3642,18 @@ class Client extends openapi_core_1.default {
|
|
|
3549
3642
|
return await this.merchandisePlacementDetectionLiteWithOptions(request, runtime);
|
|
3550
3643
|
}
|
|
3551
3644
|
/**
|
|
3552
|
-
* Detects the positions and names of visible products in an image
|
|
3645
|
+
* Detects the positions and names of visible products in an image. Built on a dedicated visual model, this product display detection service supports default recognition rules for the Genki Forest business type and allows the caller to invoke custom rules. It outputs the position and name of each product, and is applicable to beverage display inspection and product positioning.
|
|
3553
3646
|
*
|
|
3554
3647
|
* @remarks
|
|
3555
3648
|
* ## Scenarios
|
|
3556
|
-
* - Beverage display recognition: Uses default rules for the Genki Forest business type to recognize visible beverage products.
|
|
3557
|
-
* - Custom detection rules: Specifies product recognition
|
|
3558
|
-
* - Product
|
|
3649
|
+
* - Beverage display recognition: Uses the default rules for the Genki Forest business type to recognize visible beverage products.
|
|
3650
|
+
* - Custom detection rules: Specifies a product recognition task that meets output constraints by using the rule parameter.
|
|
3651
|
+
* - Product positioning display: Uses the name and bounding box of each product for display, statistics, or manual review.
|
|
3559
3652
|
* ## Features
|
|
3560
|
-
* - Direct product recognition:
|
|
3561
|
-
* - Two rule
|
|
3562
|
-
* - Unified result validation:
|
|
3563
|
-
* - Result scope: Currently outputs detected names and positions
|
|
3653
|
+
* - Direct product recognition: Outputs a bounding box and a name for each product. If the name cannot be determined, unknown is returned.
|
|
3654
|
+
* - Two rule entries: A non-empty rule takes precedence. If no valid rule is provided, the type must be Genki Forest.
|
|
3655
|
+
* - Unified result validation: Converts bounding box coordinates to integers and limits them to 0-1000, and swaps reverse coordinates. Uses unknown when the name is missing. The entire batch fails if any item is not a valid product object, the bounding box is invalid, or the bounding box has no area.
|
|
3656
|
+
* - Result scope: Currently outputs the detected names and positions, and does not include vector database recall, SKU codes, or similarity scores.
|
|
3564
3657
|
*
|
|
3565
3658
|
* @param request - MerchandisePlacementDetectionProRequest
|
|
3566
3659
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -3595,18 +3688,18 @@ class Client extends openapi_core_1.default {
|
|
|
3595
3688
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.MerchandisePlacementDetectionProResponse({}));
|
|
3596
3689
|
}
|
|
3597
3690
|
/**
|
|
3598
|
-
* Detects the positions and names of visible products in an image
|
|
3691
|
+
* Detects the positions and names of visible products in an image. Built on a dedicated visual model, this product display detection service supports default recognition rules for the Genki Forest business type and allows the caller to invoke custom rules. It outputs the position and name of each product, and is applicable to beverage display inspection and product positioning.
|
|
3599
3692
|
*
|
|
3600
3693
|
* @remarks
|
|
3601
3694
|
* ## Scenarios
|
|
3602
|
-
* - Beverage display recognition: Uses default rules for the Genki Forest business type to recognize visible beverage products.
|
|
3603
|
-
* - Custom detection rules: Specifies product recognition
|
|
3604
|
-
* - Product
|
|
3695
|
+
* - Beverage display recognition: Uses the default rules for the Genki Forest business type to recognize visible beverage products.
|
|
3696
|
+
* - Custom detection rules: Specifies a product recognition task that meets output constraints by using the rule parameter.
|
|
3697
|
+
* - Product positioning display: Uses the name and bounding box of each product for display, statistics, or manual review.
|
|
3605
3698
|
* ## Features
|
|
3606
|
-
* - Direct product recognition:
|
|
3607
|
-
* - Two rule
|
|
3608
|
-
* - Unified result validation:
|
|
3609
|
-
* - Result scope: Currently outputs detected names and positions
|
|
3699
|
+
* - Direct product recognition: Outputs a bounding box and a name for each product. If the name cannot be determined, unknown is returned.
|
|
3700
|
+
* - Two rule entries: A non-empty rule takes precedence. If no valid rule is provided, the type must be Genki Forest.
|
|
3701
|
+
* - Unified result validation: Converts bounding box coordinates to integers and limits them to 0-1000, and swaps reverse coordinates. Uses unknown when the name is missing. The entire batch fails if any item is not a valid product object, the bounding box is invalid, or the bounding box has no area.
|
|
3702
|
+
* - Result scope: Currently outputs the detected names and positions, and does not include vector database recall, SKU codes, or similarity scores.
|
|
3610
3703
|
*
|
|
3611
3704
|
* @param request - MerchandisePlacementDetectionProRequest
|
|
3612
3705
|
* @returns MerchandisePlacementDetectionProResponse
|
|
@@ -3616,18 +3709,18 @@ class Client extends openapi_core_1.default {
|
|
|
3616
3709
|
return await this.merchandisePlacementDetectionProWithOptions(request, runtime);
|
|
3617
3710
|
}
|
|
3618
3711
|
/**
|
|
3619
|
-
* Detects display price tags based on price tag recognition and multimodal large model
|
|
3712
|
+
* Detects display price tags based on price tag recognition and multimodal large model grounding. This operation extracts product information, selling prices, original prices, and locations from price tags, and provides visible bounding boxes for paper price tags, electronic price tags, and promotional price tags. It is applicable to store price collection, price tag inspection, and business-side price verification.
|
|
3620
3713
|
*
|
|
3621
3714
|
* @remarks
|
|
3622
3715
|
* ## Scenarios
|
|
3623
|
-
* - Store price collection: Reads the
|
|
3716
|
+
* - Store price collection: Reads and recognizes the selling prices and available original prices of price tags.
|
|
3624
3717
|
* - Price execution verification: Compares the recognized prices with the business-side price list.
|
|
3625
|
-
* - Manual price tag review: Locates paper, electronic, and promotional price tags
|
|
3718
|
+
* - Manual price tag review: Locates paper, electronic, and promotional price tags for easy review of the original images.
|
|
3626
3719
|
* ## Features
|
|
3627
3720
|
* - Structured price output: Returns the product name, code, price, and location item by item. Prices are retained as strings.
|
|
3628
|
-
* - Multi-type price tag
|
|
3629
|
-
* - Original text and location-assisted review: Prioritizes extracting clear selling prices
|
|
3630
|
-
* - Supports single HTTPS image input. Supplementary
|
|
3721
|
+
* - Multi-type price tag grounding: Covers ordinary paper price tags, electronic price tags, and promotional tags with clear selling prices.
|
|
3722
|
+
* - Original text and location-assisted review: Prioritizes extracting clear selling prices based on located text. Leaves the field empty if the text is unreadable. Supplementary grounding does not infer that a price tag necessarily corresponds to a specific SKU based on its location.
|
|
3723
|
+
* - Supports single HTTPS image input. Supplementary grounding returns up to 12 targets. Each target includes a category, a bounding box, and up to 24 characters of visible text, and uses the coverage field to indicate whether the grounding is complete.
|
|
3631
3724
|
*
|
|
3632
3725
|
* @param request - MerchandisePriceTagDetectionRequest
|
|
3633
3726
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -3656,18 +3749,18 @@ class Client extends openapi_core_1.default {
|
|
|
3656
3749
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.MerchandisePriceTagDetectionResponse({}));
|
|
3657
3750
|
}
|
|
3658
3751
|
/**
|
|
3659
|
-
* Detects display price tags based on price tag recognition and multimodal large model
|
|
3752
|
+
* Detects display price tags based on price tag recognition and multimodal large model grounding. This operation extracts product information, selling prices, original prices, and locations from price tags, and provides visible bounding boxes for paper price tags, electronic price tags, and promotional price tags. It is applicable to store price collection, price tag inspection, and business-side price verification.
|
|
3660
3753
|
*
|
|
3661
3754
|
* @remarks
|
|
3662
3755
|
* ## Scenarios
|
|
3663
|
-
* - Store price collection: Reads the
|
|
3756
|
+
* - Store price collection: Reads and recognizes the selling prices and available original prices of price tags.
|
|
3664
3757
|
* - Price execution verification: Compares the recognized prices with the business-side price list.
|
|
3665
|
-
* - Manual price tag review: Locates paper, electronic, and promotional price tags
|
|
3758
|
+
* - Manual price tag review: Locates paper, electronic, and promotional price tags for easy review of the original images.
|
|
3666
3759
|
* ## Features
|
|
3667
3760
|
* - Structured price output: Returns the product name, code, price, and location item by item. Prices are retained as strings.
|
|
3668
|
-
* - Multi-type price tag
|
|
3669
|
-
* - Original text and location-assisted review: Prioritizes extracting clear selling prices
|
|
3670
|
-
* - Supports single HTTPS image input. Supplementary
|
|
3761
|
+
* - Multi-type price tag grounding: Covers ordinary paper price tags, electronic price tags, and promotional tags with clear selling prices.
|
|
3762
|
+
* - Original text and location-assisted review: Prioritizes extracting clear selling prices based on located text. Leaves the field empty if the text is unreadable. Supplementary grounding does not infer that a price tag necessarily corresponds to a specific SKU based on its location.
|
|
3763
|
+
* - Supports single HTTPS image input. Supplementary grounding returns up to 12 targets. Each target includes a category, a bounding box, and up to 24 characters of visible text, and uses the coverage field to indicate whether the grounding is complete.
|
|
3671
3764
|
*
|
|
3672
3765
|
* @param request - MerchandisePriceTagDetectionRequest
|
|
3673
3766
|
* @returns MerchandisePriceTagDetectionResponse
|
|
@@ -3677,18 +3770,18 @@ class Client extends openapi_core_1.default {
|
|
|
3677
3770
|
return await this.merchandisePriceTagDetectionWithOptions(request, runtime);
|
|
3678
3771
|
}
|
|
3679
3772
|
/**
|
|
3680
|
-
* Detects store displays based on display scene recognition and holistic positioning
|
|
3773
|
+
* Detects store displays based on display scene recognition and holistic positioning of multimodal large models. Identifies display areas such as shelves, refrigerators, and floor displays, outputs available scene names, locations, and display metrics, and supplements overall facility bounding boxes for display facility inspection, area visualization, and display metric verification.
|
|
3681
3774
|
*
|
|
3682
3775
|
* @remarks
|
|
3683
3776
|
* ## Scenarios
|
|
3684
|
-
* - Display facility inspection:
|
|
3685
|
-
* - Display metric verification:
|
|
3686
|
-
* - Display area visualization:
|
|
3777
|
+
* - Display facility inspection: Identifies facilities such as shelves, endcaps, floor displays, upright refrigerators, and chest freezers.
|
|
3778
|
+
* - Display metric verification: Reads active metrics such as proportion, purity, and fullness for further evaluation by business systems.
|
|
3779
|
+
* - Display area visualization: Displays area locations and overall facility bounding boxes on the original image.
|
|
3687
3780
|
* ## Features
|
|
3688
|
-
* - Scene structured output:
|
|
3689
|
-
* - Holistic facility positioning:
|
|
3690
|
-
* - Retains two types of coordinates:
|
|
3691
|
-
* - Supports
|
|
3781
|
+
* - Scene structured output: Returns scene names, center points, area contours, and active metrics.
|
|
3782
|
+
* - Holistic facility positioning: Uses one bounding box for an independent facility or promotional display, which is suitable for overall area review.
|
|
3783
|
+
* - Retains two types of coordinates: The positions in scenes retain the original normalization coordinates of the recognition service, while grounding uses 0-1000 integer bounding boxes.
|
|
3784
|
+
* - Supports single HTTPS image input. The supplementary positioning returns up to 12 targets. Each target includes a category, a bounding box, and up to 24 characters of visible text, and uses coverage to indicate whether the positioning is complete.
|
|
3692
3785
|
*
|
|
3693
3786
|
* @param request - MerchandiseSceneDetectionRequest
|
|
3694
3787
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -3717,18 +3810,18 @@ class Client extends openapi_core_1.default {
|
|
|
3717
3810
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.MerchandiseSceneDetectionResponse({}));
|
|
3718
3811
|
}
|
|
3719
3812
|
/**
|
|
3720
|
-
* Detects store displays based on display scene recognition and holistic positioning
|
|
3813
|
+
* Detects store displays based on display scene recognition and holistic positioning of multimodal large models. Identifies display areas such as shelves, refrigerators, and floor displays, outputs available scene names, locations, and display metrics, and supplements overall facility bounding boxes for display facility inspection, area visualization, and display metric verification.
|
|
3721
3814
|
*
|
|
3722
3815
|
* @remarks
|
|
3723
3816
|
* ## Scenarios
|
|
3724
|
-
* - Display facility inspection:
|
|
3725
|
-
* - Display metric verification:
|
|
3726
|
-
* - Display area visualization:
|
|
3817
|
+
* - Display facility inspection: Identifies facilities such as shelves, endcaps, floor displays, upright refrigerators, and chest freezers.
|
|
3818
|
+
* - Display metric verification: Reads active metrics such as proportion, purity, and fullness for further evaluation by business systems.
|
|
3819
|
+
* - Display area visualization: Displays area locations and overall facility bounding boxes on the original image.
|
|
3727
3820
|
* ## Features
|
|
3728
|
-
* - Scene structured output:
|
|
3729
|
-
* - Holistic facility positioning:
|
|
3730
|
-
* - Retains two types of coordinates:
|
|
3731
|
-
* - Supports
|
|
3821
|
+
* - Scene structured output: Returns scene names, center points, area contours, and active metrics.
|
|
3822
|
+
* - Holistic facility positioning: Uses one bounding box for an independent facility or promotional display, which is suitable for overall area review.
|
|
3823
|
+
* - Retains two types of coordinates: The positions in scenes retain the original normalization coordinates of the recognition service, while grounding uses 0-1000 integer bounding boxes.
|
|
3824
|
+
* - Supports single HTTPS image input. The supplementary positioning returns up to 12 targets. Each target includes a category, a bounding box, and up to 24 characters of visible text, and uses coverage to indicate whether the positioning is complete.
|
|
3732
3825
|
*
|
|
3733
3826
|
* @param request - MerchandiseSceneDetectionRequest
|
|
3734
3827
|
* @returns MerchandiseSceneDetectionResponse
|
|
@@ -3738,16 +3831,16 @@ class Client extends openapi_core_1.default {
|
|
|
3738
3831
|
return await this.merchandiseSceneDetectionWithOptions(request, runtime);
|
|
3739
3832
|
}
|
|
3740
3833
|
/**
|
|
3741
|
-
*
|
|
3834
|
+
* Detects recaptured images of milk powder can bottom codes and localizes elements based on recapture recognition and multimodal large model capabilities. The operation determines whether an uploaded photo is recaptured and localizes visible milk powder products, can bottoms, and bottom code areas. It applies to milk powder channel inspections, bottom code photo reviews, and manual verifications.
|
|
3742
3835
|
*
|
|
3743
3836
|
* @remarks
|
|
3744
3837
|
* ## Features
|
|
3745
|
-
* - Recapture binary classification: isFake=true indicates a
|
|
3746
|
-
* - Milk powder subject and bottom code localization: Prioritizes the localization of the milk powder subject. Adds
|
|
3838
|
+
* - Recapture binary classification: isFake=true indicates a recaptured image, and false indicates a non-recaptured image. Element localization is performed for both results.
|
|
3839
|
+
* - Milk powder subject and bottom code localization: Prioritizes the localization of the milk powder subject. Adds bounding boxes for the can bottom, spray code, batch code, or traceability code areas when they are clear.
|
|
3747
3840
|
* - Visible information output: Outputs only confirmed short names or text without guessing blurred code values. Localization results do not change the recapture conclusion.
|
|
3748
|
-
* - Supports a single HTTPS image input.
|
|
3749
|
-
* ## Scenarios
|
|
3750
|
-
* -
|
|
3841
|
+
* - Supports a single HTTPS image input. Supplementary localization returns up to 12 targets. Each target includes a category, a bounding box, and up to 24 characters of visible text. The coverage field indicates whether the localization is complete.
|
|
3842
|
+
* ## Scenarios
|
|
3843
|
+
* - Bottom code photo review: Determines whether a photo of a milk powder can bottom code has a recapture risk.
|
|
3751
3844
|
* - Milk powder product inspection: Localizes canned, boxed, or bagged milk powder, as well as clearly visible can bottoms and bottom codes.
|
|
3752
3845
|
* - Review result playback: Displays target bounding boxes on the original image to assist personnel in verifying the photo content.
|
|
3753
3846
|
*
|
|
@@ -3778,16 +3871,16 @@ class Client extends openapi_core_1.default {
|
|
|
3778
3871
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.MilkPowderCanBottomCodeRephotographyDetectionResponse({}));
|
|
3779
3872
|
}
|
|
3780
3873
|
/**
|
|
3781
|
-
*
|
|
3874
|
+
* Detects recaptured images of milk powder can bottom codes and localizes elements based on recapture recognition and multimodal large model capabilities. The operation determines whether an uploaded photo is recaptured and localizes visible milk powder products, can bottoms, and bottom code areas. It applies to milk powder channel inspections, bottom code photo reviews, and manual verifications.
|
|
3782
3875
|
*
|
|
3783
3876
|
* @remarks
|
|
3784
3877
|
* ## Features
|
|
3785
|
-
* - Recapture binary classification: isFake=true indicates a
|
|
3786
|
-
* - Milk powder subject and bottom code localization: Prioritizes the localization of the milk powder subject. Adds
|
|
3878
|
+
* - Recapture binary classification: isFake=true indicates a recaptured image, and false indicates a non-recaptured image. Element localization is performed for both results.
|
|
3879
|
+
* - Milk powder subject and bottom code localization: Prioritizes the localization of the milk powder subject. Adds bounding boxes for the can bottom, spray code, batch code, or traceability code areas when they are clear.
|
|
3787
3880
|
* - Visible information output: Outputs only confirmed short names or text without guessing blurred code values. Localization results do not change the recapture conclusion.
|
|
3788
|
-
* - Supports a single HTTPS image input.
|
|
3789
|
-
* ## Scenarios
|
|
3790
|
-
* -
|
|
3881
|
+
* - Supports a single HTTPS image input. Supplementary localization returns up to 12 targets. Each target includes a category, a bounding box, and up to 24 characters of visible text. The coverage field indicates whether the localization is complete.
|
|
3882
|
+
* ## Scenarios
|
|
3883
|
+
* - Bottom code photo review: Determines whether a photo of a milk powder can bottom code has a recapture risk.
|
|
3791
3884
|
* - Milk powder product inspection: Localizes canned, boxed, or bagged milk powder, as well as clearly visible can bottoms and bottom codes.
|
|
3792
3885
|
* - Review result playback: Displays target bounding boxes on the original image to assist personnel in verifying the photo content.
|
|
3793
3886
|
*
|
|
@@ -3799,16 +3892,16 @@ class Client extends openapi_core_1.default {
|
|
|
3799
3892
|
return await this.milkPowderCanBottomCodeRephotographyDetectionWithOptions(request, runtime);
|
|
3800
3893
|
}
|
|
3801
3894
|
/**
|
|
3802
|
-
*
|
|
3895
|
+
* Analyzes milk powder box code images for channel inspection by using recapture detection and multimodal large model element grounding. Outputs non-recapture, recapture, or suspected recapture conclusions, and locates visible milk powder products, outer boxes, and code areas to provide a structured basis for photo review and manual verification.
|
|
3803
3896
|
*
|
|
3804
3897
|
* @remarks
|
|
3805
3898
|
* ## Features
|
|
3806
|
-
* - Three-state recapture detection: When photoStatus is 0, 1, or 2, it indicates non-recapture, recapture, or suspected recapture, respectively. The isFake, isSuspect, and score
|
|
3807
|
-
* - Subject and code area grounding: Prioritizes
|
|
3808
|
-
* - Result
|
|
3809
|
-
* - Supports a single HTTPS image as input. The supplementary grounding returns up to 12 targets
|
|
3899
|
+
* - Three-state recapture detection: When photoStatus is 0, 1, or 2, it indicates non-recapture, recapture, or suspected recapture, respectively. The operation also outputs isFake, isSuspect, and score.
|
|
3900
|
+
* - Subject and code area grounding: Prioritizes grounding of individual milk powder products and outer boxes, and supplements clear box codes, barcodes, and QR code areas. If no box code is present, the milk powder subject can still be grounded.
|
|
3901
|
+
* - Result synergy: Element grounding supplements visible targets without changing the recapture detection conclusion. It does not decode QR codes or provide product authenticity verification.
|
|
3902
|
+
* - Supports a single HTTPS image as input. The supplementary grounding returns up to 12 targets, each containing a category, bounding box, and up to 24 characters of visible text. The coverage field indicates whether the grounding is complete.
|
|
3810
3903
|
* ## Scenarios
|
|
3811
|
-
* - Box code photo review: Checks
|
|
3904
|
+
* - Box code photo review: Checks milk powder outer box photos uploaded by stores or channels for recapture risks.
|
|
3812
3905
|
* - Milk powder display inspection: Locates visible individual milk powder products, full boxes, and clear code areas in photos.
|
|
3813
3906
|
* - Suspected sample review: Filters images that require manual verification based on the suspected status, algorithm score, and bounding boxes.
|
|
3814
3907
|
*
|
|
@@ -3839,16 +3932,16 @@ class Client extends openapi_core_1.default {
|
|
|
3839
3932
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.MilkPowderCartonCodeRephotographyDetectionResponse({}));
|
|
3840
3933
|
}
|
|
3841
3934
|
/**
|
|
3842
|
-
*
|
|
3935
|
+
* Analyzes milk powder box code images for channel inspection by using recapture detection and multimodal large model element grounding. Outputs non-recapture, recapture, or suspected recapture conclusions, and locates visible milk powder products, outer boxes, and code areas to provide a structured basis for photo review and manual verification.
|
|
3843
3936
|
*
|
|
3844
3937
|
* @remarks
|
|
3845
3938
|
* ## Features
|
|
3846
|
-
* - Three-state recapture detection: When photoStatus is 0, 1, or 2, it indicates non-recapture, recapture, or suspected recapture, respectively. The isFake, isSuspect, and score
|
|
3847
|
-
* - Subject and code area grounding: Prioritizes
|
|
3848
|
-
* - Result
|
|
3849
|
-
* - Supports a single HTTPS image as input. The supplementary grounding returns up to 12 targets
|
|
3939
|
+
* - Three-state recapture detection: When photoStatus is 0, 1, or 2, it indicates non-recapture, recapture, or suspected recapture, respectively. The operation also outputs isFake, isSuspect, and score.
|
|
3940
|
+
* - Subject and code area grounding: Prioritizes grounding of individual milk powder products and outer boxes, and supplements clear box codes, barcodes, and QR code areas. If no box code is present, the milk powder subject can still be grounded.
|
|
3941
|
+
* - Result synergy: Element grounding supplements visible targets without changing the recapture detection conclusion. It does not decode QR codes or provide product authenticity verification.
|
|
3942
|
+
* - Supports a single HTTPS image as input. The supplementary grounding returns up to 12 targets, each containing a category, bounding box, and up to 24 characters of visible text. The coverage field indicates whether the grounding is complete.
|
|
3850
3943
|
* ## Scenarios
|
|
3851
|
-
* - Box code photo review: Checks
|
|
3944
|
+
* - Box code photo review: Checks milk powder outer box photos uploaded by stores or channels for recapture risks.
|
|
3852
3945
|
* - Milk powder display inspection: Locates visible individual milk powder products, full boxes, and clear code areas in photos.
|
|
3853
3946
|
* - Suspected sample review: Filters images that require manual verification based on the suspected status, algorithm score, and bounding boxes.
|
|
3854
3947
|
*
|
|
@@ -3982,28 +4075,28 @@ class Client extends openapi_core_1.default {
|
|
|
3982
4075
|
return await this.petHealthAnalysisWithOptions(request, runtime);
|
|
3983
4076
|
}
|
|
3984
4077
|
/**
|
|
3985
|
-
* Maas-MT
|
|
4078
|
+
* Provides Maas-MT, an e-commerce translation model trained on over 1 billion high-quality bilingual sentence pairs, which is continuously optimized for e-commerce-specific terms such as brands, models, materials, and attributes. Supports translation across more than 400 language pairs, including bridged pairs, with a focus on Chinese-English, Chinese-multilingual, and English-multilingual translations. Supports language detection for 24 language pairs and automatically detects the source language if it is not specified.
|
|
3986
4079
|
*
|
|
3987
4080
|
* @remarks
|
|
3988
|
-
* ## Product
|
|
3989
|
-
* Marco MT is an e-commerce translation model trained on over 1 billion
|
|
3990
|
-
* ##
|
|
3991
|
-
* - Product titles: Marco MT large model translation accurately
|
|
3992
|
-
* - Product descriptions: Product descriptions are
|
|
3993
|
-
* - Product attributes: Product attribute terms are
|
|
3994
|
-
* - Customer service conversations: In customer service scenarios,
|
|
3995
|
-
* Marco MT large model translation can also be used for general
|
|
4081
|
+
* ## Product introduction
|
|
4082
|
+
* Marco MT is an e-commerce translation model trained on over 1 billion high-quality bilingual sentence pairs, continuously optimized for e-commerce-specific terms such as brands, models, materials, and attributes. It supports translation across more than 100 language pairs, including bridged pairs, especially between Chinese and English, Chinese and multiple languages, and English and multiple languages. It supports language detection for 24 language pairs and automatically detects the source language if it is not specified.
|
|
4083
|
+
* ## Scenarios
|
|
4084
|
+
* - Product titles: Marco MT large model translation can accurately translate phrases, keywords, and trending words in product titles into the target language, increasing product exposure on e-commerce platforms.
|
|
4085
|
+
* - Product descriptions: Product descriptions are generally long and have diverse and complex formats. Marco MT large model translation supports long-text translation and HTML format, delivering excellent translation quality and format restoration.
|
|
4086
|
+
* - Product attributes: Product attribute terms are short and contain e-commerce-specific words, such as models and parameters.
|
|
4087
|
+
* - Customer service conversations: In customer service scenarios, it can optimize issues such as missing context and typos, and supports polite expressions in some language pairs.
|
|
4088
|
+
* Additionally, Marco MT large model translation can also be used for translation in general scenarios, such as office work, meetings, and daily conversations.
|
|
3996
4089
|
* ## Features
|
|
3997
|
-
* - **Supported language
|
|
3998
|
-
* - **Ultra-low hallucination rate and high-speed inference
|
|
3999
|
-
* - **Supported text
|
|
4000
|
-
* - **Translation intervention
|
|
4090
|
+
* - **Supported language pairs**: Marco MT large model translation supports translation between more than 100 languages and language detection for 23 languages. For specific language pairs, refer to the language mapping table in Section 5.6.
|
|
4091
|
+
* - **Ultra-low hallucination rate and high-speed inference**: Marco MT large model translation is a large language model aligned with preferences for translation scenarios. Compared with general large language models, it features faster translation inference speed and an extremely low hallucination rate, making it especially suitable for large-scale invocation scenarios.
|
|
4092
|
+
* - **Supported text formats**: Currently supports text and HTML formats.
|
|
4093
|
+
* - **Translation intervention**: Supports custom translation results, including do-not-translate (ABC-ABC), specified translation (ABC-DEF), and skip translation (ABC-null). This is commonly used for brand word protection. You only need to pass the corresponding intervention glossary ID when calling the API. This meets your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention terms. If you need more, contact the platform.
|
|
4001
4094
|
* ## Translation quality comparison
|
|
4002
|
-
* | **Source text** | **Aidge translation** | **Other translation product 1** | **Other translation product 2** | **Aidge translation
|
|
4095
|
+
* | **Source text** | **Aidge translation** | **Other translation product 1** | **Other translation product 2** | **Aidge translation advantages** |
|
|
4003
4096
|
* |----------|---------------|-----------------------|-----------------------|---------------------|
|
|
4004
4097
|
* | 75mm 4Pin 0.38A GA81S2U 12V GTX 960 Cooling Fan for Zotac GTX960 Graphics Card Cooling Fan | 75mm 4Pin 0.38A GA81S2U 12V GTX 960, ventilador de enfriamiento para tarjeta gráfica Zotac GTX960. | Ventilador de refrigeración de 75mm 4Pin 0.38A GA81S2U 12V para la tarjeta gráfica Zotac GTX960 Ventilador de enfriamiento GTX 960. | Ventilador de refrigeración para tarjeta gráfica Zotac GTX960, 75mm, 4 pines, 0,38A, GA81S2U, 12V, GTX 960 | Clearer product subject |
|
|
4005
|
-
* | New Arrival Classic Style Luxury Brand 6 Hands Swiss Automatic Mechanical Men\\"s Business Steel Watch | Nueva llegada, reloj mecánico automático de lujo con 6 manecillas, estilo clásico, de acero para hombres de negocios | Nueva Llegada Reloj De Acero de Negocios para Hombre con Mecanismo Automático Suizo y 6 Agujas de Estilo Clásico de Marca de Lujo. | Reloj de acero de negocios para hombre, mecánico automático suizo, marca de lujo, estilo clásico, 6 manos, nueva llegada | More complete sentence structure, clearer expression, and more reasonable
|
|
4006
|
-
* | Men Key Bag Genuine Cow Leather Buckets Key Cases Pouch Zipper Keychain Auto Car Key Case Women Home Key Holder Wallet | Sac à clés pour hommes en cuir de vache véritable, étuis à clés, pochette à fermeture éclair, porte-clés, étui à clé de voiture, porte-clés pour femmes, organisateur de portefeuille | Porte-clés sac en cuir de vache véritable, étuis à clés en forme de seau, pochette à fermeture éclair, porte-clés auto pour voiture, étui à clés pour femmes, organisateur de clés de maison, portefeuille. | Hommes porte-clés en cuir de vache véritable sac/pochette fermature éclair porte-clés Auto voiture porte-clés femmes maison porte-clés organisateur portefeuille | Clearer and more
|
|
4098
|
+
* | New Arrival Classic Style Luxury Brand 6 Hands Swiss Automatic Mechanical Men\\"s Business Steel Watch | Nueva llegada, reloj mecánico automático de lujo con 6 manecillas, estilo clásico, de acero para hombres de negocios | Nueva Llegada Reloj De Acero de Negocios para Hombre con Mecanismo Automático Suizo y 6 Agujas de Estilo Clásico de Marca de Lujo. | Reloj de acero de negocios para hombre, mecánico automático suizo, marca de lujo, estilo clásico, 6 manos, nueva llegada | More complete sentence structure, clearer expression, and more reasonable punctuation |
|
|
4099
|
+
* | Men Key Bag Genuine Cow Leather Buckets Key Cases Pouch Zipper Keychain Auto Car Key Case Women Home Key Holder Wallet | Sac à clés pour hommes en cuir de vache véritable, étuis à clés, pochette à fermeture éclair, porte-clés, étui à clé de voiture, porte-clés pour femmes, organisateur de portefeuille | Porte-clés sac en cuir de vache véritable, étuis à clés en forme de seau, pochette à fermeture éclair, porte-clés auto pour voiture, étui à clés pour femmes, organisateur de clés de maison, portefeuille. | Hommes porte-clés en cuir de vache véritable sac/pochette fermature éclair porte-clés Auto voiture porte-clés femmes maison porte-clés organisateur portefeuille | Clearer and more explicit product subject expression, identifying it as a "key bag" rather than a "keychain", whereas other translations may cause ambiguity regarding the product subject |
|
|
4007
4100
|
*
|
|
4008
4101
|
* @param tmpReq - PrepaidTextTranslateRequest
|
|
4009
4102
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -4055,28 +4148,28 @@ class Client extends openapi_core_1.default {
|
|
|
4055
4148
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.PrepaidTextTranslateResponse({}));
|
|
4056
4149
|
}
|
|
4057
4150
|
/**
|
|
4058
|
-
* Maas-MT
|
|
4151
|
+
* Provides Maas-MT, an e-commerce translation model trained on over 1 billion high-quality bilingual sentence pairs, which is continuously optimized for e-commerce-specific terms such as brands, models, materials, and attributes. Supports translation across more than 400 language pairs, including bridged pairs, with a focus on Chinese-English, Chinese-multilingual, and English-multilingual translations. Supports language detection for 24 language pairs and automatically detects the source language if it is not specified.
|
|
4059
4152
|
*
|
|
4060
4153
|
* @remarks
|
|
4061
|
-
* ## Product
|
|
4062
|
-
* Marco MT is an e-commerce translation model trained on over 1 billion
|
|
4063
|
-
* ##
|
|
4064
|
-
* - Product titles: Marco MT large model translation accurately
|
|
4065
|
-
* - Product descriptions: Product descriptions are
|
|
4066
|
-
* - Product attributes: Product attribute terms are
|
|
4067
|
-
* - Customer service conversations: In customer service scenarios,
|
|
4068
|
-
* Marco MT large model translation can also be used for general
|
|
4154
|
+
* ## Product introduction
|
|
4155
|
+
* Marco MT is an e-commerce translation model trained on over 1 billion high-quality bilingual sentence pairs, continuously optimized for e-commerce-specific terms such as brands, models, materials, and attributes. It supports translation across more than 100 language pairs, including bridged pairs, especially between Chinese and English, Chinese and multiple languages, and English and multiple languages. It supports language detection for 24 language pairs and automatically detects the source language if it is not specified.
|
|
4156
|
+
* ## Scenarios
|
|
4157
|
+
* - Product titles: Marco MT large model translation can accurately translate phrases, keywords, and trending words in product titles into the target language, increasing product exposure on e-commerce platforms.
|
|
4158
|
+
* - Product descriptions: Product descriptions are generally long and have diverse and complex formats. Marco MT large model translation supports long-text translation and HTML format, delivering excellent translation quality and format restoration.
|
|
4159
|
+
* - Product attributes: Product attribute terms are short and contain e-commerce-specific words, such as models and parameters.
|
|
4160
|
+
* - Customer service conversations: In customer service scenarios, it can optimize issues such as missing context and typos, and supports polite expressions in some language pairs.
|
|
4161
|
+
* Additionally, Marco MT large model translation can also be used for translation in general scenarios, such as office work, meetings, and daily conversations.
|
|
4069
4162
|
* ## Features
|
|
4070
|
-
* - **Supported language
|
|
4071
|
-
* - **Ultra-low hallucination rate and high-speed inference
|
|
4072
|
-
* - **Supported text
|
|
4073
|
-
* - **Translation intervention
|
|
4163
|
+
* - **Supported language pairs**: Marco MT large model translation supports translation between more than 100 languages and language detection for 23 languages. For specific language pairs, refer to the language mapping table in Section 5.6.
|
|
4164
|
+
* - **Ultra-low hallucination rate and high-speed inference**: Marco MT large model translation is a large language model aligned with preferences for translation scenarios. Compared with general large language models, it features faster translation inference speed and an extremely low hallucination rate, making it especially suitable for large-scale invocation scenarios.
|
|
4165
|
+
* - **Supported text formats**: Currently supports text and HTML formats.
|
|
4166
|
+
* - **Translation intervention**: Supports custom translation results, including do-not-translate (ABC-ABC), specified translation (ABC-DEF), and skip translation (ABC-null). This is commonly used for brand word protection. You only need to pass the corresponding intervention glossary ID when calling the API. This meets your translation needs in different scenarios. Generally, you can upload up to 100,000 intervention terms. If you need more, contact the platform.
|
|
4074
4167
|
* ## Translation quality comparison
|
|
4075
|
-
* | **Source text** | **Aidge translation** | **Other translation product 1** | **Other translation product 2** | **Aidge translation
|
|
4168
|
+
* | **Source text** | **Aidge translation** | **Other translation product 1** | **Other translation product 2** | **Aidge translation advantages** |
|
|
4076
4169
|
* |----------|---------------|-----------------------|-----------------------|---------------------|
|
|
4077
4170
|
* | 75mm 4Pin 0.38A GA81S2U 12V GTX 960 Cooling Fan for Zotac GTX960 Graphics Card Cooling Fan | 75mm 4Pin 0.38A GA81S2U 12V GTX 960, ventilador de enfriamiento para tarjeta gráfica Zotac GTX960. | Ventilador de refrigeración de 75mm 4Pin 0.38A GA81S2U 12V para la tarjeta gráfica Zotac GTX960 Ventilador de enfriamiento GTX 960. | Ventilador de refrigeración para tarjeta gráfica Zotac GTX960, 75mm, 4 pines, 0,38A, GA81S2U, 12V, GTX 960 | Clearer product subject |
|
|
4078
|
-
* | New Arrival Classic Style Luxury Brand 6 Hands Swiss Automatic Mechanical Men\\"s Business Steel Watch | Nueva llegada, reloj mecánico automático de lujo con 6 manecillas, estilo clásico, de acero para hombres de negocios | Nueva Llegada Reloj De Acero de Negocios para Hombre con Mecanismo Automático Suizo y 6 Agujas de Estilo Clásico de Marca de Lujo. | Reloj de acero de negocios para hombre, mecánico automático suizo, marca de lujo, estilo clásico, 6 manos, nueva llegada | More complete sentence structure, clearer expression, and more reasonable
|
|
4079
|
-
* | Men Key Bag Genuine Cow Leather Buckets Key Cases Pouch Zipper Keychain Auto Car Key Case Women Home Key Holder Wallet | Sac à clés pour hommes en cuir de vache véritable, étuis à clés, pochette à fermeture éclair, porte-clés, étui à clé de voiture, porte-clés pour femmes, organisateur de portefeuille | Porte-clés sac en cuir de vache véritable, étuis à clés en forme de seau, pochette à fermeture éclair, porte-clés auto pour voiture, étui à clés pour femmes, organisateur de clés de maison, portefeuille. | Hommes porte-clés en cuir de vache véritable sac/pochette fermature éclair porte-clés Auto voiture porte-clés femmes maison porte-clés organisateur portefeuille | Clearer and more
|
|
4171
|
+
* | New Arrival Classic Style Luxury Brand 6 Hands Swiss Automatic Mechanical Men\\"s Business Steel Watch | Nueva llegada, reloj mecánico automático de lujo con 6 manecillas, estilo clásico, de acero para hombres de negocios | Nueva Llegada Reloj De Acero de Negocios para Hombre con Mecanismo Automático Suizo y 6 Agujas de Estilo Clásico de Marca de Lujo. | Reloj de acero de negocios para hombre, mecánico automático suizo, marca de lujo, estilo clásico, 6 manos, nueva llegada | More complete sentence structure, clearer expression, and more reasonable punctuation |
|
|
4172
|
+
* | Men Key Bag Genuine Cow Leather Buckets Key Cases Pouch Zipper Keychain Auto Car Key Case Women Home Key Holder Wallet | Sac à clés pour hommes en cuir de vache véritable, étuis à clés, pochette à fermeture éclair, porte-clés, étui à clé de voiture, porte-clés pour femmes, organisateur de portefeuille | Porte-clés sac en cuir de vache véritable, étuis à clés en forme de seau, pochette à fermeture éclair, porte-clés auto pour voiture, étui à clés pour femmes, organisateur de clés de maison, portefeuille. | Hommes porte-clés en cuir de vache véritable sac/pochette fermature éclair porte-clés Auto voiture porte-clés femmes maison porte-clés organisateur portefeuille | Clearer and more explicit product subject expression, identifying it as a "key bag" rather than a "keychain", whereas other translations may cause ambiguity regarding the product subject |
|
|
4080
4173
|
*
|
|
4081
4174
|
* @param request - PrepaidTextTranslateRequest
|
|
4082
4175
|
* @returns PrepaidTextTranslateResponse
|
|
@@ -4157,6 +4250,95 @@ class Client extends openapi_core_1.default {
|
|
|
4157
4250
|
let runtime = new $dara.RuntimeOptions({});
|
|
4158
4251
|
return await this.productHotspotDetectionWithOptions(request, runtime);
|
|
4159
4252
|
}
|
|
4253
|
+
/**
|
|
4254
|
+
* Generates a single-product showcase video for e-commerce merchants and platforms based on the product title and multiple images of the same product. This API optionally accepts a storyboard script, automatically selects reference images, plans a three-shot storyboard, and generates the video. It uses asynchronous tasks, supports 720p and 1080p resolutions and four aspect ratios, and is suitable for content production of non-apparel products. For apparel products, use the E-commerce Video Enhanced Edition.
|
|
4255
|
+
*
|
|
4256
|
+
* @remarks
|
|
4257
|
+
* # Product introduction
|
|
4258
|
+
* The basic e-commerce video edition is designed for e-commerce merchants and platforms. It generates a single-product showcase video based on the product title and multiple images of the same product. You can optionally provide a storyboard script. The system automatically selects reference images, plans a three-shot storyboard, and generates the video. The API uses asynchronous tasks, supports 720p and 1080p resolutions and four aspect ratios, and is suitable for content production of non-apparel products. For apparel products, use the e-commerce enhanced edition.
|
|
4259
|
+
* # Scenarios
|
|
4260
|
+
* - New product material generation: Provide the product title and images to automatically plan storyboard shots such as product debut, details, or usage demonstrations.
|
|
4261
|
+
* - Video production with existing scripts: Provide a storyboard script or creative description. The system optimizes the storyboard script based on the product images and retains shot intentions that are supported by the images.
|
|
4262
|
+
* - Adaptation to different video placements: Use the aspect ratio parameter to adapt to vertical short videos, main product images, detail pages, and horizontal displays. For batch production, each request corresponds to one product and one video.
|
|
4263
|
+
* # Features
|
|
4264
|
+
* - Automatic image selection and storyboard planning: When no script is provided, the system generates a storyboard based on the product title and images. When a script is provided, the system still performs image selection and planning, eliminating the need for callers to organize video prompts.
|
|
4265
|
+
* - Product appearance consistency constraint: The system uses the selected complete product image as the appearance baseline. It checks the color scheme, structure, proportions, texture, version, and specifications of other images, and selects only reference images that can be confirmed to belong to the same SKU.
|
|
4266
|
+
* - Flexible video specifications: Without a script, the duration is 5 to 15 seconds. With a script, the duration is 5 to 30 seconds. The system supports four aspect ratios (16:9, 1:1, 3:4, and 9:16) and two resolutions (720p and 1080p).
|
|
4267
|
+
* - Asynchronous generation and task query: After a task is created, a task ID is returned. The caller queries the result by using the query API. The final video includes background music, no voiceover, and no additional subtitles or marketing text.
|
|
4268
|
+
* # Examples
|
|
4269
|
+
* [View the attachment "Digital_040_WiFi6_Gigabit_Router.mp4" in the DingTalk document.](https://alidocs.dingtalk.com/i/nodes/gwva2dxOW4vRkd9DUn1g1p6kJbkz3BRL?cid=6760411388:7016395022&corpId=dingd8e1123006514592&iframeQuery=anchorId%3DX02mvsyycduj3nepwoeush&utm_medium=im_card&utm_scene=person_space&utm_source=im)
|
|
4270
|
+
*
|
|
4271
|
+
* @param tmpReq - ProductVideoGenerationRequest
|
|
4272
|
+
* @param runtime - runtime options for this request RuntimeOptions
|
|
4273
|
+
* @returns ProductVideoGenerationResponse
|
|
4274
|
+
*/
|
|
4275
|
+
async productVideoGenerationWithOptions(tmpReq, runtime) {
|
|
4276
|
+
tmpReq.validate();
|
|
4277
|
+
let request = new $_model.ProductVideoGenerationShrinkRequest({});
|
|
4278
|
+
openapi_core_2.OpenApiUtil.convert(tmpReq, request);
|
|
4279
|
+
if (!$dara.isNull(tmpReq.productImages)) {
|
|
4280
|
+
request.productImagesShrink = openapi_core_2.OpenApiUtil.arrayToStringWithSpecifiedStyle(tmpReq.productImages, "product_images", "json");
|
|
4281
|
+
}
|
|
4282
|
+
let query = {};
|
|
4283
|
+
if (!$dara.isNull(request.aspectRatio)) {
|
|
4284
|
+
query["aspect_ratio"] = request.aspectRatio;
|
|
4285
|
+
}
|
|
4286
|
+
if (!$dara.isNull(request.duration)) {
|
|
4287
|
+
query["duration"] = request.duration;
|
|
4288
|
+
}
|
|
4289
|
+
if (!$dara.isNull(request.productImagesShrink)) {
|
|
4290
|
+
query["product_images"] = request.productImagesShrink;
|
|
4291
|
+
}
|
|
4292
|
+
if (!$dara.isNull(request.productTitle)) {
|
|
4293
|
+
query["product_title"] = request.productTitle;
|
|
4294
|
+
}
|
|
4295
|
+
if (!$dara.isNull(request.resolution)) {
|
|
4296
|
+
query["resolution"] = request.resolution;
|
|
4297
|
+
}
|
|
4298
|
+
if (!$dara.isNull(request.script)) {
|
|
4299
|
+
query["script"] = request.script;
|
|
4300
|
+
}
|
|
4301
|
+
let req = new openapi_core_2.$OpenApiUtil.OpenApiRequest({
|
|
4302
|
+
query: openapi_core_2.OpenApiUtil.query(query),
|
|
4303
|
+
});
|
|
4304
|
+
let params = new openapi_core_2.$OpenApiUtil.Params({
|
|
4305
|
+
action: "ProductVideoGeneration",
|
|
4306
|
+
version: "2026-04-28",
|
|
4307
|
+
protocol: "HTTPS",
|
|
4308
|
+
pathname: "/",
|
|
4309
|
+
method: "POST",
|
|
4310
|
+
authType: "AK",
|
|
4311
|
+
style: "RPC",
|
|
4312
|
+
reqBodyType: "formData",
|
|
4313
|
+
bodyType: "json",
|
|
4314
|
+
});
|
|
4315
|
+
return $dara.cast(await this.callApi(params, req, runtime), new $_model.ProductVideoGenerationResponse({}));
|
|
4316
|
+
}
|
|
4317
|
+
/**
|
|
4318
|
+
* Generates a single-product showcase video for e-commerce merchants and platforms based on the product title and multiple images of the same product. This API optionally accepts a storyboard script, automatically selects reference images, plans a three-shot storyboard, and generates the video. It uses asynchronous tasks, supports 720p and 1080p resolutions and four aspect ratios, and is suitable for content production of non-apparel products. For apparel products, use the E-commerce Video Enhanced Edition.
|
|
4319
|
+
*
|
|
4320
|
+
* @remarks
|
|
4321
|
+
* # Product introduction
|
|
4322
|
+
* The basic e-commerce video edition is designed for e-commerce merchants and platforms. It generates a single-product showcase video based on the product title and multiple images of the same product. You can optionally provide a storyboard script. The system automatically selects reference images, plans a three-shot storyboard, and generates the video. The API uses asynchronous tasks, supports 720p and 1080p resolutions and four aspect ratios, and is suitable for content production of non-apparel products. For apparel products, use the e-commerce enhanced edition.
|
|
4323
|
+
* # Scenarios
|
|
4324
|
+
* - New product material generation: Provide the product title and images to automatically plan storyboard shots such as product debut, details, or usage demonstrations.
|
|
4325
|
+
* - Video production with existing scripts: Provide a storyboard script or creative description. The system optimizes the storyboard script based on the product images and retains shot intentions that are supported by the images.
|
|
4326
|
+
* - Adaptation to different video placements: Use the aspect ratio parameter to adapt to vertical short videos, main product images, detail pages, and horizontal displays. For batch production, each request corresponds to one product and one video.
|
|
4327
|
+
* # Features
|
|
4328
|
+
* - Automatic image selection and storyboard planning: When no script is provided, the system generates a storyboard based on the product title and images. When a script is provided, the system still performs image selection and planning, eliminating the need for callers to organize video prompts.
|
|
4329
|
+
* - Product appearance consistency constraint: The system uses the selected complete product image as the appearance baseline. It checks the color scheme, structure, proportions, texture, version, and specifications of other images, and selects only reference images that can be confirmed to belong to the same SKU.
|
|
4330
|
+
* - Flexible video specifications: Without a script, the duration is 5 to 15 seconds. With a script, the duration is 5 to 30 seconds. The system supports four aspect ratios (16:9, 1:1, 3:4, and 9:16) and two resolutions (720p and 1080p).
|
|
4331
|
+
* - Asynchronous generation and task query: After a task is created, a task ID is returned. The caller queries the result by using the query API. The final video includes background music, no voiceover, and no additional subtitles or marketing text.
|
|
4332
|
+
* # Examples
|
|
4333
|
+
* [View the attachment "Digital_040_WiFi6_Gigabit_Router.mp4" in the DingTalk document.](https://alidocs.dingtalk.com/i/nodes/gwva2dxOW4vRkd9DUn1g1p6kJbkz3BRL?cid=6760411388:7016395022&corpId=dingd8e1123006514592&iframeQuery=anchorId%3DX02mvsyycduj3nepwoeush&utm_medium=im_card&utm_scene=person_space&utm_source=im)
|
|
4334
|
+
*
|
|
4335
|
+
* @param request - ProductVideoGenerationRequest
|
|
4336
|
+
* @returns ProductVideoGenerationResponse
|
|
4337
|
+
*/
|
|
4338
|
+
async productVideoGeneration(request) {
|
|
4339
|
+
let runtime = new $dara.RuntimeOptions({});
|
|
4340
|
+
return await this.productVideoGenerationWithOptions(request, runtime);
|
|
4341
|
+
}
|
|
4160
4342
|
/**
|
|
4161
4343
|
* Queries the result of an asynchronous task. The API name in the POP gateway is QueryAsyncTaskResult.
|
|
4162
4344
|
*
|
|
@@ -4622,105 +4804,105 @@ class Client extends openapi_core_1.default {
|
|
|
4622
4804
|
return await this.sizeChartExtractWithOptions(request, runtime);
|
|
4623
4805
|
}
|
|
4624
4806
|
/**
|
|
4625
|
-
* Extracts structured information such as size charts and
|
|
4807
|
+
* Extracts structured information such as size charts and package weight and dimensions from images by using a multimodal large language model in asynchronous mode.
|
|
4626
4808
|
*
|
|
4627
4809
|
* @remarks
|
|
4628
4810
|
* ## Scenarios
|
|
4629
4811
|
* Size charts: Extracts size charts of apparel categories into text.
|
|
4630
|
-
*
|
|
4812
|
+
* Package weight and dimensions: Extracts package weight and dimension information from images.
|
|
4631
4813
|
* ## Features
|
|
4632
|
-
* The size chart extraction service is based on a multimodal large model. It automatically recognizes size charts or
|
|
4633
|
-
* Core capabilities: You input an image URL and optional column name configurations, and the system outputs the structured data of the size chart or
|
|
4634
|
-
* ## Comparison
|
|
4814
|
+
* The size chart extraction service is based on a multimodal large language model. It automatically recognizes size charts or package weight and dimension information in images and converts the information into structured text output. You provide an image URL, and the system processes the request as an asynchronous task and returns the extracted structured data. You can specify the column names to be recognized, such as Size, Bust, and Length, by using ColumnNameList to achieve precise and targeted extraction. You can also control the output language by using the LanguageModel parameter. Valid values: cn for Chinese and en for English. If you do not specify this parameter, the original language is retained. This service is applicable to scenarios such as text conversion of size charts for apparel categories and batch structuring of package weight and dimension information for commodities.
|
|
4815
|
+
* Core capabilities: You input an image URL and optional column name configurations, and the system outputs the structured data of the size chart or package weight and dimensions in the image. The API operation uses the asynchronous task mode (background: true). After you submit the task, a task ID is returned. You must poll the task to obtain the final result. This mode is suitable for processing complex images or handling large-scale calls.
|
|
4816
|
+
* ## Comparison of results
|
|
4635
4817
|
* | Input image | Output result |
|
|
4636
4818
|
* | --- | --- |
|
|
4637
4819
|
* |  | {
|
|
4638
4820
|
* "SizeChart": [
|
|
4639
4821
|
* {
|
|
4640
|
-
* "
|
|
4822
|
+
* "Size": {
|
|
4641
4823
|
* "value": "S"
|
|
4642
4824
|
* },
|
|
4643
|
-
* "
|
|
4825
|
+
* "Waist": {
|
|
4644
4826
|
* "unit": "CM",
|
|
4645
4827
|
* "value": 58
|
|
4646
4828
|
* },
|
|
4647
|
-
* "
|
|
4829
|
+
* "Skirt length": {
|
|
4648
4830
|
* "unit": "CM",
|
|
4649
4831
|
* "value": 81
|
|
4650
4832
|
* },
|
|
4651
|
-
* "
|
|
4833
|
+
* "Recommended weight": {
|
|
4652
4834
|
* "unit": "kg",
|
|
4653
4835
|
* "lower_value": 40,
|
|
4654
4836
|
* "upper_value": 47.5
|
|
4655
4837
|
* }
|
|
4656
4838
|
* },
|
|
4657
4839
|
* {
|
|
4658
|
-
* "
|
|
4840
|
+
* "Size": {
|
|
4659
4841
|
* "value": "M"
|
|
4660
4842
|
* },
|
|
4661
|
-
* "
|
|
4843
|
+
* "Waist": {
|
|
4662
4844
|
* "unit": "CM",
|
|
4663
4845
|
* "value": 59
|
|
4664
4846
|
* },
|
|
4665
|
-
* "
|
|
4847
|
+
* "Skirt length": {
|
|
4666
4848
|
* "unit": "CM",
|
|
4667
4849
|
* "value": 82
|
|
4668
4850
|
* },
|
|
4669
|
-
* "
|
|
4851
|
+
* "Recommended weight": {
|
|
4670
4852
|
* "unit": "kg",
|
|
4671
4853
|
* "lower_value": 45,
|
|
4672
4854
|
* "upper_value": 52.5
|
|
4673
4855
|
* }
|
|
4674
4856
|
* },
|
|
4675
4857
|
* {
|
|
4676
|
-
* "
|
|
4858
|
+
* "Size": {
|
|
4677
4859
|
* "value": "L"
|
|
4678
4860
|
* },
|
|
4679
|
-
* "
|
|
4861
|
+
* "Waist": {
|
|
4680
4862
|
* "unit": "CM",
|
|
4681
4863
|
* "value": 60
|
|
4682
4864
|
* },
|
|
4683
|
-
* "
|
|
4865
|
+
* "Skirt length": {
|
|
4684
4866
|
* "unit": "CM",
|
|
4685
4867
|
* "value": 83
|
|
4686
4868
|
* },
|
|
4687
|
-
* "
|
|
4869
|
+
* "Recommended weight": {
|
|
4688
4870
|
* "unit": "kg",
|
|
4689
4871
|
* "lower_value": 50,
|
|
4690
4872
|
* "upper_value": 57.5
|
|
4691
4873
|
* }
|
|
4692
4874
|
* },
|
|
4693
4875
|
* {
|
|
4694
|
-
* "
|
|
4876
|
+
* "Size": {
|
|
4695
4877
|
* "value": "XL"
|
|
4696
4878
|
* },
|
|
4697
|
-
* "
|
|
4879
|
+
* "Waist": {
|
|
4698
4880
|
* "unit": "CM",
|
|
4699
4881
|
* "value": 61
|
|
4700
4882
|
* },
|
|
4701
|
-
* "
|
|
4883
|
+
* "Skirt length": {
|
|
4702
4884
|
* "unit": "CM",
|
|
4703
4885
|
* "value": 84
|
|
4704
4886
|
* },
|
|
4705
|
-
* "
|
|
4887
|
+
* "Recommended weight": {
|
|
4706
4888
|
* "unit": "kg",
|
|
4707
4889
|
* "lower_value": 55,
|
|
4708
4890
|
* "upper_value": 62.5
|
|
4709
4891
|
* }
|
|
4710
4892
|
* },
|
|
4711
4893
|
* {
|
|
4712
|
-
* "
|
|
4894
|
+
* "Size": {
|
|
4713
4895
|
* "value": "2XL"
|
|
4714
4896
|
* },
|
|
4715
|
-
* "
|
|
4897
|
+
* "Waist": {
|
|
4716
4898
|
* "unit": "CM",
|
|
4717
4899
|
* "value": 62
|
|
4718
4900
|
* },
|
|
4719
|
-
* "
|
|
4901
|
+
* "Skirt length": {
|
|
4720
4902
|
* "unit": "CM",
|
|
4721
4903
|
* "value": 85
|
|
4722
4904
|
* },
|
|
4723
|
-
* "
|
|
4905
|
+
* "Recommended weight": {
|
|
4724
4906
|
* "unit": "kg",
|
|
4725
4907
|
* "lower_value": 60,
|
|
4726
4908
|
* "upper_value": 67.5
|
|
@@ -4731,13 +4913,13 @@ class Client extends openapi_core_1.default {
|
|
|
4731
4913
|
* "Size",
|
|
4732
4914
|
* "Skirt length",
|
|
4733
4915
|
* "Waist",
|
|
4734
|
-
* "
|
|
4916
|
+
* "Recommended weight"
|
|
4735
4917
|
* \\],
|
|
4736
4918
|
* "IntersectionAttributes": [
|
|
4737
4919
|
* "Size",
|
|
4738
4920
|
* "Skirt length",
|
|
4739
|
-
* "Waist (
|
|
4740
|
-
* "
|
|
4921
|
+
* "Waist (stretched)",
|
|
4922
|
+
* "Recommended weight"
|
|
4741
4923
|
* \\]
|
|
4742
4924
|
* } |
|
|
4743
4925
|
*
|
|
@@ -4782,105 +4964,105 @@ class Client extends openapi_core_1.default {
|
|
|
4782
4964
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.SizeChartExtractAsyncResponse({}));
|
|
4783
4965
|
}
|
|
4784
4966
|
/**
|
|
4785
|
-
* Extracts structured information such as size charts and
|
|
4967
|
+
* Extracts structured information such as size charts and package weight and dimensions from images by using a multimodal large language model in asynchronous mode.
|
|
4786
4968
|
*
|
|
4787
4969
|
* @remarks
|
|
4788
4970
|
* ## Scenarios
|
|
4789
4971
|
* Size charts: Extracts size charts of apparel categories into text.
|
|
4790
|
-
*
|
|
4972
|
+
* Package weight and dimensions: Extracts package weight and dimension information from images.
|
|
4791
4973
|
* ## Features
|
|
4792
|
-
* The size chart extraction service is based on a multimodal large model. It automatically recognizes size charts or
|
|
4793
|
-
* Core capabilities: You input an image URL and optional column name configurations, and the system outputs the structured data of the size chart or
|
|
4794
|
-
* ## Comparison
|
|
4974
|
+
* The size chart extraction service is based on a multimodal large language model. It automatically recognizes size charts or package weight and dimension information in images and converts the information into structured text output. You provide an image URL, and the system processes the request as an asynchronous task and returns the extracted structured data. You can specify the column names to be recognized, such as Size, Bust, and Length, by using ColumnNameList to achieve precise and targeted extraction. You can also control the output language by using the LanguageModel parameter. Valid values: cn for Chinese and en for English. If you do not specify this parameter, the original language is retained. This service is applicable to scenarios such as text conversion of size charts for apparel categories and batch structuring of package weight and dimension information for commodities.
|
|
4975
|
+
* Core capabilities: You input an image URL and optional column name configurations, and the system outputs the structured data of the size chart or package weight and dimensions in the image. The API operation uses the asynchronous task mode (background: true). After you submit the task, a task ID is returned. You must poll the task to obtain the final result. This mode is suitable for processing complex images or handling large-scale calls.
|
|
4976
|
+
* ## Comparison of results
|
|
4795
4977
|
* | Input image | Output result |
|
|
4796
4978
|
* | --- | --- |
|
|
4797
4979
|
* |  | {
|
|
4798
4980
|
* "SizeChart": [
|
|
4799
4981
|
* {
|
|
4800
|
-
* "
|
|
4982
|
+
* "Size": {
|
|
4801
4983
|
* "value": "S"
|
|
4802
4984
|
* },
|
|
4803
|
-
* "
|
|
4985
|
+
* "Waist": {
|
|
4804
4986
|
* "unit": "CM",
|
|
4805
4987
|
* "value": 58
|
|
4806
4988
|
* },
|
|
4807
|
-
* "
|
|
4989
|
+
* "Skirt length": {
|
|
4808
4990
|
* "unit": "CM",
|
|
4809
4991
|
* "value": 81
|
|
4810
4992
|
* },
|
|
4811
|
-
* "
|
|
4993
|
+
* "Recommended weight": {
|
|
4812
4994
|
* "unit": "kg",
|
|
4813
4995
|
* "lower_value": 40,
|
|
4814
4996
|
* "upper_value": 47.5
|
|
4815
4997
|
* }
|
|
4816
4998
|
* },
|
|
4817
4999
|
* {
|
|
4818
|
-
* "
|
|
5000
|
+
* "Size": {
|
|
4819
5001
|
* "value": "M"
|
|
4820
5002
|
* },
|
|
4821
|
-
* "
|
|
5003
|
+
* "Waist": {
|
|
4822
5004
|
* "unit": "CM",
|
|
4823
5005
|
* "value": 59
|
|
4824
5006
|
* },
|
|
4825
|
-
* "
|
|
5007
|
+
* "Skirt length": {
|
|
4826
5008
|
* "unit": "CM",
|
|
4827
5009
|
* "value": 82
|
|
4828
5010
|
* },
|
|
4829
|
-
* "
|
|
5011
|
+
* "Recommended weight": {
|
|
4830
5012
|
* "unit": "kg",
|
|
4831
5013
|
* "lower_value": 45,
|
|
4832
5014
|
* "upper_value": 52.5
|
|
4833
5015
|
* }
|
|
4834
5016
|
* },
|
|
4835
5017
|
* {
|
|
4836
|
-
* "
|
|
5018
|
+
* "Size": {
|
|
4837
5019
|
* "value": "L"
|
|
4838
5020
|
* },
|
|
4839
|
-
* "
|
|
5021
|
+
* "Waist": {
|
|
4840
5022
|
* "unit": "CM",
|
|
4841
5023
|
* "value": 60
|
|
4842
5024
|
* },
|
|
4843
|
-
* "
|
|
5025
|
+
* "Skirt length": {
|
|
4844
5026
|
* "unit": "CM",
|
|
4845
5027
|
* "value": 83
|
|
4846
5028
|
* },
|
|
4847
|
-
* "
|
|
5029
|
+
* "Recommended weight": {
|
|
4848
5030
|
* "unit": "kg",
|
|
4849
5031
|
* "lower_value": 50,
|
|
4850
5032
|
* "upper_value": 57.5
|
|
4851
5033
|
* }
|
|
4852
5034
|
* },
|
|
4853
5035
|
* {
|
|
4854
|
-
* "
|
|
5036
|
+
* "Size": {
|
|
4855
5037
|
* "value": "XL"
|
|
4856
5038
|
* },
|
|
4857
|
-
* "
|
|
5039
|
+
* "Waist": {
|
|
4858
5040
|
* "unit": "CM",
|
|
4859
5041
|
* "value": 61
|
|
4860
5042
|
* },
|
|
4861
|
-
* "
|
|
5043
|
+
* "Skirt length": {
|
|
4862
5044
|
* "unit": "CM",
|
|
4863
5045
|
* "value": 84
|
|
4864
5046
|
* },
|
|
4865
|
-
* "
|
|
5047
|
+
* "Recommended weight": {
|
|
4866
5048
|
* "unit": "kg",
|
|
4867
5049
|
* "lower_value": 55,
|
|
4868
5050
|
* "upper_value": 62.5
|
|
4869
5051
|
* }
|
|
4870
5052
|
* },
|
|
4871
5053
|
* {
|
|
4872
|
-
* "
|
|
5054
|
+
* "Size": {
|
|
4873
5055
|
* "value": "2XL"
|
|
4874
5056
|
* },
|
|
4875
|
-
* "
|
|
5057
|
+
* "Waist": {
|
|
4876
5058
|
* "unit": "CM",
|
|
4877
5059
|
* "value": 62
|
|
4878
5060
|
* },
|
|
4879
|
-
* "
|
|
5061
|
+
* "Skirt length": {
|
|
4880
5062
|
* "unit": "CM",
|
|
4881
5063
|
* "value": 85
|
|
4882
5064
|
* },
|
|
4883
|
-
* "
|
|
5065
|
+
* "Recommended weight": {
|
|
4884
5066
|
* "unit": "kg",
|
|
4885
5067
|
* "lower_value": 60,
|
|
4886
5068
|
* "upper_value": 67.5
|
|
@@ -4891,13 +5073,13 @@ class Client extends openapi_core_1.default {
|
|
|
4891
5073
|
* "Size",
|
|
4892
5074
|
* "Skirt length",
|
|
4893
5075
|
* "Waist",
|
|
4894
|
-
* "
|
|
5076
|
+
* "Recommended weight"
|
|
4895
5077
|
* \\],
|
|
4896
5078
|
* "IntersectionAttributes": [
|
|
4897
5079
|
* "Size",
|
|
4898
5080
|
* "Skirt length",
|
|
4899
|
-
* "Waist (
|
|
4900
|
-
* "
|
|
5081
|
+
* "Waist (stretched)",
|
|
5082
|
+
* "Recommended weight"
|
|
4901
5083
|
* \\]
|
|
4902
5084
|
* } |
|
|
4903
5085
|
*
|
|
@@ -4909,18 +5091,18 @@ class Client extends openapi_core_1.default {
|
|
|
4909
5091
|
return await this.sizeChartExtractAsyncWithOptions(request, runtime);
|
|
4910
5092
|
}
|
|
4911
5093
|
/**
|
|
4912
|
-
*
|
|
5094
|
+
* Audits store photos based on storefront scene recognition and multimodal large model element localization. The API outputs storefront scene classifications and confidence details, and localizes storefronts, signboards, entrances, and business signs. It is applicable to store data collection, storefront photo classification, and inspection reviews.
|
|
4913
5095
|
*
|
|
4914
5096
|
* @remarks
|
|
4915
|
-
* ##
|
|
5097
|
+
* ## Scenarios
|
|
4916
5098
|
* - Storefront photo classification: Distinguishes storefront-related photos from non-storefront photos.
|
|
4917
5099
|
* - Store data verification: Localizes signboards and entrances, and extracts clear and readable short text.
|
|
4918
5100
|
* - Multi-store photo review: Localizes visible storefronts and signs of adjacent stores separately.
|
|
4919
5101
|
* ## Features
|
|
4920
|
-
* - Scene classification and details: Returns
|
|
5102
|
+
* - Scene classification and details: Returns a scenes string array and classification details. Non-storefront photos are also valid classification results.
|
|
4921
5103
|
* - Storefront key element localization: Covers storefronts, signboards, store entrances, and business-related signs.
|
|
4922
|
-
* - Readable text extraction: Extracts only clear short text.
|
|
4923
|
-
* - Supports single HTTPS image input. Supplementary localization returns up to 12 targets, each containing a category,
|
|
5104
|
+
* - Readable text extraction: Extracts only clear short text. Does not infer business status based on whether the door is open or closed, and does not infer addresses or store ownership.
|
|
5105
|
+
* - Supports single HTTPS image input. Supplementary localization returns up to 12 targets, each containing a category, bounding box, and up to 24 characters of visible text, with coverage indicating whether the localization is complete.
|
|
4924
5106
|
*
|
|
4925
5107
|
* @param request - StorefrontSceneRecognitionRequest
|
|
4926
5108
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -4949,18 +5131,18 @@ class Client extends openapi_core_1.default {
|
|
|
4949
5131
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.StorefrontSceneRecognitionResponse({}));
|
|
4950
5132
|
}
|
|
4951
5133
|
/**
|
|
4952
|
-
*
|
|
5134
|
+
* Audits store photos based on storefront scene recognition and multimodal large model element localization. The API outputs storefront scene classifications and confidence details, and localizes storefronts, signboards, entrances, and business signs. It is applicable to store data collection, storefront photo classification, and inspection reviews.
|
|
4953
5135
|
*
|
|
4954
5136
|
* @remarks
|
|
4955
|
-
* ##
|
|
5137
|
+
* ## Scenarios
|
|
4956
5138
|
* - Storefront photo classification: Distinguishes storefront-related photos from non-storefront photos.
|
|
4957
5139
|
* - Store data verification: Localizes signboards and entrances, and extracts clear and readable short text.
|
|
4958
5140
|
* - Multi-store photo review: Localizes visible storefronts and signs of adjacent stores separately.
|
|
4959
5141
|
* ## Features
|
|
4960
|
-
* - Scene classification and details: Returns
|
|
5142
|
+
* - Scene classification and details: Returns a scenes string array and classification details. Non-storefront photos are also valid classification results.
|
|
4961
5143
|
* - Storefront key element localization: Covers storefronts, signboards, store entrances, and business-related signs.
|
|
4962
|
-
* - Readable text extraction: Extracts only clear short text.
|
|
4963
|
-
* - Supports single HTTPS image input. Supplementary localization returns up to 12 targets, each containing a category,
|
|
5144
|
+
* - Readable text extraction: Extracts only clear short text. Does not infer business status based on whether the door is open or closed, and does not infer addresses or store ownership.
|
|
5145
|
+
* - Supports single HTTPS image input. Supplementary localization returns up to 12 targets, each containing a category, bounding box, and up to 24 characters of visible text, with coverage indicating whether the localization is complete.
|
|
4964
5146
|
*
|
|
4965
5147
|
* @param request - StorefrontSceneRecognitionRequest
|
|
4966
5148
|
* @returns StorefrontSceneRecognitionResponse
|
|
@@ -4970,7 +5152,7 @@ class Client extends openapi_core_1.default {
|
|
|
4970
5152
|
return await this.storefrontSceneRecognitionWithOptions(request, runtime);
|
|
4971
5153
|
}
|
|
4972
5154
|
/**
|
|
4973
|
-
* Generates product detail page
|
|
5155
|
+
* Generates product detail page materials automatically based on a multimodal large language model (MLLM) and structured image-text extraction to help e-commerce platforms and brands supplement product detail pages. After the caller pushes product information such as the product ID, title, brand, main image, and shop name, the service automatically locates identical products, performs consistency checks, extracts size information and product attributes, renders them into SVG size and attribute charts for direct use on the detail page, and returns structured detail copy classifications and processing remarks.
|
|
4974
5156
|
*
|
|
4975
5157
|
* @param request - SubmitProductMatchRequest
|
|
4976
5158
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -5017,7 +5199,7 @@ class Client extends openapi_core_1.default {
|
|
|
5017
5199
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.SubmitProductMatchResponse({}));
|
|
5018
5200
|
}
|
|
5019
5201
|
/**
|
|
5020
|
-
* Generates product detail page
|
|
5202
|
+
* Generates product detail page materials automatically based on a multimodal large language model (MLLM) and structured image-text extraction to help e-commerce platforms and brands supplement product detail pages. After the caller pushes product information such as the product ID, title, brand, main image, and shop name, the service automatically locates identical products, performs consistency checks, extracts size information and product attributes, renders them into SVG size and attribute charts for direct use on the detail page, and returns structured detail copy classifications and processing remarks.
|
|
5021
5203
|
*
|
|
5022
5204
|
* @param request - SubmitProductMatchRequest
|
|
5023
5205
|
* @returns SubmitProductMatchResponse
|
|
@@ -5225,7 +5407,7 @@ class Client extends openapi_core_1.default {
|
|
|
5225
5407
|
return await this.textTranslateWithOptions(request, runtime);
|
|
5226
5408
|
}
|
|
5227
5409
|
/**
|
|
5228
|
-
* Translates text by using the Maas MT e-commerce translation model
|
|
5410
|
+
* Translates text asynchronously by using the Maas MT model, an e-commerce translation model trained on over 1 billion high-quality bilingual sentence pairs. The model is continuously optimized for e-commerce-specific terms such as brands, models, materials, and attributes. It supports 156 languages with over 400 language pairs and supports automatic language detection for source languages.
|
|
5229
5411
|
*
|
|
5230
5412
|
* @remarks
|
|
5231
5413
|
* ## Scenarios
|
|
@@ -5239,22 +5421,22 @@ class Client extends openapi_core_1.default {
|
|
|
5239
5421
|
*
|
|
5240
5422
|
* The Maas MT large language model can also be used for translation in general scenarios, such as office work, meetings, and daily conversations.
|
|
5241
5423
|
* ## Features
|
|
5242
|
-
* * **Ultra-low hallucination rate and high-speed inference
|
|
5424
|
+
* * **Ultra-low hallucination rate and high-speed inference** The Maas MT large language model is aligned with preferences for translation scenarios. Compared with general-purpose large language models, it features faster translation inference and an extremely low hallucination rate, making it especially suitable for large-scale invocation scenarios.
|
|
5243
5425
|
*
|
|
5244
|
-
* * **Supported text formats
|
|
5426
|
+
* * **Supported text formats** The model supports text and HTML formats.
|
|
5245
5427
|
*
|
|
5246
|
-
* * **Translation intervention
|
|
5428
|
+
* * **Translation intervention** You can customize translation results, including keeping words untranslated (ABC-ABC), specifying translations (ABC-DEF), and omitting translations (ABC-null value). This is commonly used for brand word protection. You only need to pass the corresponding intervention glossary ID when calling the API. This meets your translation needs in different scenarios. You can upload up to 100,000 intervention words. If you need more, contact the platform.
|
|
5247
5429
|
*
|
|
5248
|
-
* * **Translation
|
|
5430
|
+
* * **Translation quality comparison**
|
|
5249
5431
|
* | **Source text** | **Aidge translation** | **Other translation product 1** | **Other translation product 2** | **Aidge translation advantages** |
|
|
5250
5432
|
* | --- | --- | --- | --- | --- |
|
|
5251
5433
|
* | 75mm 4Pin 0.38A GA81S2U 12V GTX 960 Cooling Fan for Zotac GTX960 Graphics Card Cooling Fan | 75mm 4Pin 0.38A GA81S2U 12V GTX 960, ventilador de enfriamiento para tarjeta gráfica Zotac GTX960. | Ventilador de refrigeración de 75mm 4Pin 0.38A GA81S2U 12V para la tarjeta gráfica Zotac GTX960 Ventilador de enfriamiento GTX 960. | Ventilador de refrigeración para tarjeta gráfica Zotac GTX960, 75mm, 4 pines, 0,38A, GA81S2U, 12V, GTX 960 | Clearer product subject |
|
|
5252
5434
|
* | New Arrival Classic Style Luxury Brand 6 Hands Swiss Automatic Mechanical Men\\"s Business Steel Watch | Nueva llegada, reloj mecánico automático de lujo con 6 manecillas, estilo clásico, de acero para hombres de negocios | Nueva Llegada Reloj De Acero de Negocios para Hombre con Mecanismo Automático Suizo y 6 Agujas de Estilo Clásico de Marca de Lujo. | Reloj de acero de negocios para hombre, mecánico automático suizo, marca de lujo, estilo clásico, 6 manos, nueva llegada | More complete sentence structure, clearer expression, and more reasonable punctuation |
|
|
5253
|
-
* | Men Key Bag Genuine Cow Leather Buckets Key Cases Pouch Zipper Keychain Auto Car Key Case Women Home Key Holder Wallet | Sac à clés pour hommes en cuir de vache véritable, étuis à clés, pochette à fermeture éclair, porte-clés, étui à clé de voiture, porte-clés pour femmes, organisateur de portefeuille | Porte-clés sac en cuir de vache véritable, étuis à clés en forme de seau, pochette à fermeture éclair, porte-clés auto pour voiture, étui à clés pour femmes, organisateur de clés de maison, portefeuille. | Hommes porte-clés en cuir de vache véritable sac/pochette
|
|
5254
|
-
* ## Preparations and
|
|
5255
|
-
* Text translation supports both synchronization and asynchronous
|
|
5435
|
+
* | Men Key Bag Genuine Cow Leather Buckets Key Cases Pouch Zipper Keychain Auto Car Key Case Women Home Key Holder Wallet | Sac à clés pour hommes en cuir de vache véritable, étuis à clés, pochette à fermeture éclair, porte-clés, étui à clé de voiture, porte-clés pour femmes, organisateur de portefeuille | Porte-clés sac en cuir de vache véritable, étuis à clés en forme de seau, pochette à fermeture éclair, porte-clés auto pour voiture, étui à clés pour femmes, organisateur de clés de maison, portefeuille. | Hommes porte-clés en cuir de vache véritable sac/pochette fermeture éclair porte-clés Auto voiture porte-clés femmes maison porte-clés organisateur portefeuille | Clearer and more explicit product subject. It translates to "key bag" rather than "keychain", avoiding ambiguity in other translations |
|
|
5436
|
+
* ## Preparations and invoke flow
|
|
5437
|
+
* Text translation supports both synchronization and asynchronous invoke methods. This topic describes the asynchronous invoke API. If you need to use synchronous call, see [Text translation - Synchronous API](https://api.aliyun.com/document/Aidge/2026-04-28/TextTranslate).
|
|
5256
5438
|
* The asynchronous APIs for text translation and other capabilities share the same result retrieval flow, RocketMQ subscribe rules, message encryption protocols, dependency configurations, and Java consumption and decryption code.
|
|
5257
|
-
* For the asynchronous
|
|
5439
|
+
* For the asynchronous invoke integration flow, subscribe rules, message formats and encryption protocols, and detailed integration solutions, see [Asynchronous API invoke instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
5258
5440
|
*
|
|
5259
5441
|
* @param tmpReq - TextTranslateAsyncRequest
|
|
5260
5442
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -5309,7 +5491,7 @@ class Client extends openapi_core_1.default {
|
|
|
5309
5491
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.TextTranslateAsyncResponse({}));
|
|
5310
5492
|
}
|
|
5311
5493
|
/**
|
|
5312
|
-
* Translates text by using the Maas MT e-commerce translation model
|
|
5494
|
+
* Translates text asynchronously by using the Maas MT model, an e-commerce translation model trained on over 1 billion high-quality bilingual sentence pairs. The model is continuously optimized for e-commerce-specific terms such as brands, models, materials, and attributes. It supports 156 languages with over 400 language pairs and supports automatic language detection for source languages.
|
|
5313
5495
|
*
|
|
5314
5496
|
* @remarks
|
|
5315
5497
|
* ## Scenarios
|
|
@@ -5323,22 +5505,22 @@ class Client extends openapi_core_1.default {
|
|
|
5323
5505
|
*
|
|
5324
5506
|
* The Maas MT large language model can also be used for translation in general scenarios, such as office work, meetings, and daily conversations.
|
|
5325
5507
|
* ## Features
|
|
5326
|
-
* * **Ultra-low hallucination rate and high-speed inference
|
|
5508
|
+
* * **Ultra-low hallucination rate and high-speed inference** The Maas MT large language model is aligned with preferences for translation scenarios. Compared with general-purpose large language models, it features faster translation inference and an extremely low hallucination rate, making it especially suitable for large-scale invocation scenarios.
|
|
5327
5509
|
*
|
|
5328
|
-
* * **Supported text formats
|
|
5510
|
+
* * **Supported text formats** The model supports text and HTML formats.
|
|
5329
5511
|
*
|
|
5330
|
-
* * **Translation intervention
|
|
5512
|
+
* * **Translation intervention** You can customize translation results, including keeping words untranslated (ABC-ABC), specifying translations (ABC-DEF), and omitting translations (ABC-null value). This is commonly used for brand word protection. You only need to pass the corresponding intervention glossary ID when calling the API. This meets your translation needs in different scenarios. You can upload up to 100,000 intervention words. If you need more, contact the platform.
|
|
5331
5513
|
*
|
|
5332
|
-
* * **Translation
|
|
5514
|
+
* * **Translation quality comparison**
|
|
5333
5515
|
* | **Source text** | **Aidge translation** | **Other translation product 1** | **Other translation product 2** | **Aidge translation advantages** |
|
|
5334
5516
|
* | --- | --- | --- | --- | --- |
|
|
5335
5517
|
* | 75mm 4Pin 0.38A GA81S2U 12V GTX 960 Cooling Fan for Zotac GTX960 Graphics Card Cooling Fan | 75mm 4Pin 0.38A GA81S2U 12V GTX 960, ventilador de enfriamiento para tarjeta gráfica Zotac GTX960. | Ventilador de refrigeración de 75mm 4Pin 0.38A GA81S2U 12V para la tarjeta gráfica Zotac GTX960 Ventilador de enfriamiento GTX 960. | Ventilador de refrigeración para tarjeta gráfica Zotac GTX960, 75mm, 4 pines, 0,38A, GA81S2U, 12V, GTX 960 | Clearer product subject |
|
|
5336
5518
|
* | New Arrival Classic Style Luxury Brand 6 Hands Swiss Automatic Mechanical Men\\"s Business Steel Watch | Nueva llegada, reloj mecánico automático de lujo con 6 manecillas, estilo clásico, de acero para hombres de negocios | Nueva Llegada Reloj De Acero de Negocios para Hombre con Mecanismo Automático Suizo y 6 Agujas de Estilo Clásico de Marca de Lujo. | Reloj de acero de negocios para hombre, mecánico automático suizo, marca de lujo, estilo clásico, 6 manos, nueva llegada | More complete sentence structure, clearer expression, and more reasonable punctuation |
|
|
5337
|
-
* | Men Key Bag Genuine Cow Leather Buckets Key Cases Pouch Zipper Keychain Auto Car Key Case Women Home Key Holder Wallet | Sac à clés pour hommes en cuir de vache véritable, étuis à clés, pochette à fermeture éclair, porte-clés, étui à clé de voiture, porte-clés pour femmes, organisateur de portefeuille | Porte-clés sac en cuir de vache véritable, étuis à clés en forme de seau, pochette à fermeture éclair, porte-clés auto pour voiture, étui à clés pour femmes, organisateur de clés de maison, portefeuille. | Hommes porte-clés en cuir de vache véritable sac/pochette
|
|
5338
|
-
* ## Preparations and
|
|
5339
|
-
* Text translation supports both synchronization and asynchronous
|
|
5519
|
+
* | Men Key Bag Genuine Cow Leather Buckets Key Cases Pouch Zipper Keychain Auto Car Key Case Women Home Key Holder Wallet | Sac à clés pour hommes en cuir de vache véritable, étuis à clés, pochette à fermeture éclair, porte-clés, étui à clé de voiture, porte-clés pour femmes, organisateur de portefeuille | Porte-clés sac en cuir de vache véritable, étuis à clés en forme de seau, pochette à fermeture éclair, porte-clés auto pour voiture, étui à clés pour femmes, organisateur de clés de maison, portefeuille. | Hommes porte-clés en cuir de vache véritable sac/pochette fermeture éclair porte-clés Auto voiture porte-clés femmes maison porte-clés organisateur portefeuille | Clearer and more explicit product subject. It translates to "key bag" rather than "keychain", avoiding ambiguity in other translations |
|
|
5520
|
+
* ## Preparations and invoke flow
|
|
5521
|
+
* Text translation supports both synchronization and asynchronous invoke methods. This topic describes the asynchronous invoke API. If you need to use synchronous call, see [Text translation - Synchronous API](https://api.aliyun.com/document/Aidge/2026-04-28/TextTranslate).
|
|
5340
5522
|
* The asynchronous APIs for text translation and other capabilities share the same result retrieval flow, RocketMQ subscribe rules, message encryption protocols, dependency configurations, and Java consumption and decryption code.
|
|
5341
|
-
* For the asynchronous
|
|
5523
|
+
* For the asynchronous invoke integration flow, subscribe rules, message formats and encryption protocols, and detailed integration solutions, see [Asynchronous API invoke instructions](https://www.alibabacloud.com/help/en/aidge/asynchronous-interface-call-instructions).
|
|
5342
5524
|
*
|
|
5343
5525
|
* @param request - TextTranslateAsyncRequest
|
|
5344
5526
|
* @returns TextTranslateAsyncResponse
|
|
@@ -5395,25 +5577,70 @@ class Client extends openapi_core_1.default {
|
|
|
5395
5577
|
return await this.updateDisplayKnowledgeWithOptions(request, runtime);
|
|
5396
5578
|
}
|
|
5397
5579
|
/**
|
|
5398
|
-
*
|
|
5580
|
+
* Detects vacancies (asynchronous commit).
|
|
5581
|
+
*
|
|
5582
|
+
* @param request - VacancyDetectionRequest
|
|
5583
|
+
* @param runtime - runtime options for this request RuntimeOptions
|
|
5584
|
+
* @returns VacancyDetectionResponse
|
|
5585
|
+
*/
|
|
5586
|
+
async vacancyDetectionWithOptions(request, runtime) {
|
|
5587
|
+
request.validate();
|
|
5588
|
+
let query = {};
|
|
5589
|
+
if (!$dara.isNull(request.imageUrl)) {
|
|
5590
|
+
query["ImageUrl"] = request.imageUrl;
|
|
5591
|
+
}
|
|
5592
|
+
if (!$dara.isNull(request.systemPrompt)) {
|
|
5593
|
+
query["SystemPrompt"] = request.systemPrompt;
|
|
5594
|
+
}
|
|
5595
|
+
if (!$dara.isNull(request.userPrompt)) {
|
|
5596
|
+
query["UserPrompt"] = request.userPrompt;
|
|
5597
|
+
}
|
|
5598
|
+
let req = new openapi_core_2.$OpenApiUtil.OpenApiRequest({
|
|
5599
|
+
query: openapi_core_2.OpenApiUtil.query(query),
|
|
5600
|
+
});
|
|
5601
|
+
let params = new openapi_core_2.$OpenApiUtil.Params({
|
|
5602
|
+
action: "VacancyDetection",
|
|
5603
|
+
version: "2026-04-28",
|
|
5604
|
+
protocol: "HTTPS",
|
|
5605
|
+
pathname: "/",
|
|
5606
|
+
method: "POST",
|
|
5607
|
+
authType: "AK",
|
|
5608
|
+
style: "RPC",
|
|
5609
|
+
reqBodyType: "formData",
|
|
5610
|
+
bodyType: "json",
|
|
5611
|
+
});
|
|
5612
|
+
return $dara.cast(await this.callApi(params, req, runtime), new $_model.VacancyDetectionResponse({}));
|
|
5613
|
+
}
|
|
5614
|
+
/**
|
|
5615
|
+
* Detects vacancies (asynchronous commit).
|
|
5616
|
+
*
|
|
5617
|
+
* @param request - VacancyDetectionRequest
|
|
5618
|
+
* @returns VacancyDetectionResponse
|
|
5619
|
+
*/
|
|
5620
|
+
async vacancyDetection(request) {
|
|
5621
|
+
let runtime = new $dara.RuntimeOptions({});
|
|
5622
|
+
return await this.vacancyDetectionWithOptions(request, runtime);
|
|
5623
|
+
}
|
|
5624
|
+
/**
|
|
5625
|
+
* Generates 5- to 30-second product showcase short videos through the e-commerce AI video generation agent for e-commerce merchants and platforms based on product main images, titles, and selling points, helping customers efficiently produce and distribute e-commerce content. The product supports multiple resolutions such as 720P and 1080P, and allows you to invoke single products quickly and generate multiple products in batches, meeting the needs of daily operations and large-scale product content production. Current capabilities overwrite all clothing categories (tops, pants, skirts, suits, underwear, swimwear, etc.), and the service is gradually expanding to core e-commerce categories such as skincare, cosmetics, food and beverages, and home appliances.
|
|
5399
5626
|
*
|
|
5400
5627
|
* @remarks
|
|
5401
5628
|
* # Scenarios
|
|
5402
|
-
* - Batch generation of product short videos for sellers and platforms: Supports multiple video aspect ratios and sizes (3:4, 9:16), adapting to product detail page video slots, product homepage displays, information feed placements, short video promotions, and event venue materials.
|
|
5403
|
-
* - Quick generation of the first batch of video materials for new product launches: When merchants have a large number of new products to launch, they only need to upload basic information such as product images and titles to quickly generate product showcase videos, model wearing videos, and
|
|
5404
|
-
* - High-fidelity clothing display: Generates model display videos that are closer to real commercial shooting effects
|
|
5405
|
-
* - Generation of high-quality commercial shooting models: Built-in target audience model library, covering plus-size women, plus-size men, swimwear,
|
|
5629
|
+
* - Batch generation of product short videos for sellers and platforms: Supports multiple video aspect ratios and sizes (3:4, 9:16), adapting to scenarios such as product detail page video slots, product homepage displays, information feed placements, short video promotions, and event venue materials.
|
|
5630
|
+
* - Quick generation of the first batch of video materials for new product launches: When merchants have a large number of new products to launch at once, they only need to upload basic information such as product images and titles to quickly generate product showcase videos, model wearing videos, and voiceover explanation videos, helping merchants quickly complete material deployment during the launch phase.
|
|
5631
|
+
* - High-fidelity clothing display: Generates model display videos based on product main images that are closer to real commercial shooting effects, focusing on optimizing capabilities such as pattern restoration, fabric texture retention, accurate color presentation, and stable and consistent details. It supports common e-commerce display methods such as wearing, turning, walking, close-up shots, and multi-angle switching.
|
|
5632
|
+
* - Generation of high-quality commercial shooting models: Built-in target audience model library, covering plus-size women, plus-size men, swimwear, girly style, light mature commuting, youth casual, business menswear, sports and fitness, and other body types, skin colors, and races. It supports various commercial shooting expressions such as standing display, side display, walking display, and partial close-ups.
|
|
5406
5633
|
* - High-frequency e-commerce scenarios such as standard product explanations: Applicable to standardized product categories such as skincare, cosmetics, personal care small appliances, food and beverages, home cleaning daily necessities, and 3C digital products. It can quickly generate product explanation videos based on product main images, titles, selling point information, and explanation scripts.
|
|
5407
|
-
* - Multi-product and multi-selling point testing to improve delivery efficiency: For information feed advertising, short video
|
|
5634
|
+
* - Multi-product and multi-selling point testing to improve delivery efficiency: For information feed advertising, short video placement, and content promotion scenarios, it can quickly generate multiple versions of video materials, making it easy for enterprises to verify the conversion effects of different selling points and presentation forms at a low cost, and improve delivery ROI.
|
|
5408
5635
|
* # Features
|
|
5409
|
-
* - Fully automatic end-to-end batch generation with zero prompt threshold: Callers only need to pass in the product image URL and title, and the system automatically completes the entire process from product understanding
|
|
5410
|
-
* - Highly stable generation with a low waste rate: Adopts a deterministic generation process and multi-layer quality control mechanism to output usable materials with fewer retries, reducing invalid generation and resource waste, and making the overall production cost more controllable.
|
|
5411
|
-
* - E-commerce exclusive product fidelity capability: Built-in multi-round quality inspection
|
|
5412
|
-
* - Reference image first + 6-axis consistency check: Adopts the Reference-Image-First generation paradigm, using the original product image as the only visual source of truth, and automatically performs 6-dimensional VL diagnosis (color drift, pattern missing, contour deviation, structure addition
|
|
5413
|
-
* - Flexible storyboard arrangement and duration control: Supports adjustable video duration of 5 to 30 seconds
|
|
5414
|
-
* - Cinematic camera movement capabilities: Built-in advanced camera movement modes such as bullet time, Grammy slow motion, 360° surround rotation, and
|
|
5636
|
+
* - Fully automatic end-to-end batch generation with zero prompt threshold: Callers only need to pass in the product image URL and title, and the system automatically completes the entire process from product understanding and storyboard planning to video generation, without the need to manually write prompts, select templates, or repeatedly adjust parameters. Built-in multiple agents such as product understanding, image quality inspection, script planning, and director camera movement greatly lower the threshold for use.
|
|
5637
|
+
* - Highly stable generation with a low waste rate: Adopts a deterministic generation process and a multi-layer quality control mechanism to output usable materials with fewer retries, reducing invalid generation and resource waste, and making the overall production cost more controllable.
|
|
5638
|
+
* - E-commerce exclusive product fidelity capability: Built-in multi-round quality inspection agents lock the product state throughout the video generation process (such as zippers will not be pulled open, sleeves will not be rolled down, and prints will not disappear), ensuring product consistency through multi-layer defense of script rewriting, explicit constraint injection, and failure-aware retry.
|
|
5639
|
+
* - Reference image first + 6-axis consistency check: Adopts the Reference-Image-First generation paradigm, using the original product image as the only visual source of truth, and automatically performs 6-dimensional VL diagnosis (color drift, pattern missing, contour deviation, structure addition or reduction, model matching, and dirt detection) to fundamentally eliminate the problem of generated images not looking like the product.
|
|
5640
|
+
* - Flexible storyboard arrangement and duration control: Supports adjustable video duration of 5 to 30 seconds. Agents independently plan independent storyboards to display the product effects in different scenarios. It also supports fixed scene mode to deeply display product tonality under a unified visual style.
|
|
5641
|
+
* - Cinematic camera movement capabilities: Built-in advanced camera movement modes such as bullet time, Grammy slow motion, 360° surround rotation, and full-body model rotation display, achieving deterministic output through precise segmented control.
|
|
5415
5642
|
* - Model library and customization services: Built-in multi-type model resource library, supporting merchants to specify fixed models, customize exclusive model images, and bind brand visual personas for a long time to ensure brand visual consistency.
|
|
5416
|
-
* -
|
|
5643
|
+
* - Adaptation to high-frequency e-commerce content scenarios: Supports multiple resolutions such as 720P and 1080P and multiple aspect ratios such as 3:4 and 9:16, adapting to product detail page video slots, product homepage video displays, waterfall or information feed placements, event venue materials, and new product and promotional video generation.
|
|
5417
5644
|
*
|
|
5418
5645
|
* @param tmpReq - VideoGenerationRequest
|
|
5419
5646
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -5459,25 +5686,25 @@ class Client extends openapi_core_1.default {
|
|
|
5459
5686
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.VideoGenerationResponse({}));
|
|
5460
5687
|
}
|
|
5461
5688
|
/**
|
|
5462
|
-
*
|
|
5689
|
+
* Generates 5- to 30-second product showcase short videos through the e-commerce AI video generation agent for e-commerce merchants and platforms based on product main images, titles, and selling points, helping customers efficiently produce and distribute e-commerce content. The product supports multiple resolutions such as 720P and 1080P, and allows you to invoke single products quickly and generate multiple products in batches, meeting the needs of daily operations and large-scale product content production. Current capabilities overwrite all clothing categories (tops, pants, skirts, suits, underwear, swimwear, etc.), and the service is gradually expanding to core e-commerce categories such as skincare, cosmetics, food and beverages, and home appliances.
|
|
5463
5690
|
*
|
|
5464
5691
|
* @remarks
|
|
5465
5692
|
* # Scenarios
|
|
5466
|
-
* - Batch generation of product short videos for sellers and platforms: Supports multiple video aspect ratios and sizes (3:4, 9:16), adapting to product detail page video slots, product homepage displays, information feed placements, short video promotions, and event venue materials.
|
|
5467
|
-
* - Quick generation of the first batch of video materials for new product launches: When merchants have a large number of new products to launch, they only need to upload basic information such as product images and titles to quickly generate product showcase videos, model wearing videos, and
|
|
5468
|
-
* - High-fidelity clothing display: Generates model display videos that are closer to real commercial shooting effects
|
|
5469
|
-
* - Generation of high-quality commercial shooting models: Built-in target audience model library, covering plus-size women, plus-size men, swimwear,
|
|
5693
|
+
* - Batch generation of product short videos for sellers and platforms: Supports multiple video aspect ratios and sizes (3:4, 9:16), adapting to scenarios such as product detail page video slots, product homepage displays, information feed placements, short video promotions, and event venue materials.
|
|
5694
|
+
* - Quick generation of the first batch of video materials for new product launches: When merchants have a large number of new products to launch at once, they only need to upload basic information such as product images and titles to quickly generate product showcase videos, model wearing videos, and voiceover explanation videos, helping merchants quickly complete material deployment during the launch phase.
|
|
5695
|
+
* - High-fidelity clothing display: Generates model display videos based on product main images that are closer to real commercial shooting effects, focusing on optimizing capabilities such as pattern restoration, fabric texture retention, accurate color presentation, and stable and consistent details. It supports common e-commerce display methods such as wearing, turning, walking, close-up shots, and multi-angle switching.
|
|
5696
|
+
* - Generation of high-quality commercial shooting models: Built-in target audience model library, covering plus-size women, plus-size men, swimwear, girly style, light mature commuting, youth casual, business menswear, sports and fitness, and other body types, skin colors, and races. It supports various commercial shooting expressions such as standing display, side display, walking display, and partial close-ups.
|
|
5470
5697
|
* - High-frequency e-commerce scenarios such as standard product explanations: Applicable to standardized product categories such as skincare, cosmetics, personal care small appliances, food and beverages, home cleaning daily necessities, and 3C digital products. It can quickly generate product explanation videos based on product main images, titles, selling point information, and explanation scripts.
|
|
5471
|
-
* - Multi-product and multi-selling point testing to improve delivery efficiency: For information feed advertising, short video
|
|
5698
|
+
* - Multi-product and multi-selling point testing to improve delivery efficiency: For information feed advertising, short video placement, and content promotion scenarios, it can quickly generate multiple versions of video materials, making it easy for enterprises to verify the conversion effects of different selling points and presentation forms at a low cost, and improve delivery ROI.
|
|
5472
5699
|
* # Features
|
|
5473
|
-
* - Fully automatic end-to-end batch generation with zero prompt threshold: Callers only need to pass in the product image URL and title, and the system automatically completes the entire process from product understanding
|
|
5474
|
-
* - Highly stable generation with a low waste rate: Adopts a deterministic generation process and multi-layer quality control mechanism to output usable materials with fewer retries, reducing invalid generation and resource waste, and making the overall production cost more controllable.
|
|
5475
|
-
* - E-commerce exclusive product fidelity capability: Built-in multi-round quality inspection
|
|
5476
|
-
* - Reference image first + 6-axis consistency check: Adopts the Reference-Image-First generation paradigm, using the original product image as the only visual source of truth, and automatically performs 6-dimensional VL diagnosis (color drift, pattern missing, contour deviation, structure addition
|
|
5477
|
-
* - Flexible storyboard arrangement and duration control: Supports adjustable video duration of 5 to 30 seconds
|
|
5478
|
-
* - Cinematic camera movement capabilities: Built-in advanced camera movement modes such as bullet time, Grammy slow motion, 360° surround rotation, and
|
|
5700
|
+
* - Fully automatic end-to-end batch generation with zero prompt threshold: Callers only need to pass in the product image URL and title, and the system automatically completes the entire process from product understanding and storyboard planning to video generation, without the need to manually write prompts, select templates, or repeatedly adjust parameters. Built-in multiple agents such as product understanding, image quality inspection, script planning, and director camera movement greatly lower the threshold for use.
|
|
5701
|
+
* - Highly stable generation with a low waste rate: Adopts a deterministic generation process and a multi-layer quality control mechanism to output usable materials with fewer retries, reducing invalid generation and resource waste, and making the overall production cost more controllable.
|
|
5702
|
+
* - E-commerce exclusive product fidelity capability: Built-in multi-round quality inspection agents lock the product state throughout the video generation process (such as zippers will not be pulled open, sleeves will not be rolled down, and prints will not disappear), ensuring product consistency through multi-layer defense of script rewriting, explicit constraint injection, and failure-aware retry.
|
|
5703
|
+
* - Reference image first + 6-axis consistency check: Adopts the Reference-Image-First generation paradigm, using the original product image as the only visual source of truth, and automatically performs 6-dimensional VL diagnosis (color drift, pattern missing, contour deviation, structure addition or reduction, model matching, and dirt detection) to fundamentally eliminate the problem of generated images not looking like the product.
|
|
5704
|
+
* - Flexible storyboard arrangement and duration control: Supports adjustable video duration of 5 to 30 seconds. Agents independently plan independent storyboards to display the product effects in different scenarios. It also supports fixed scene mode to deeply display product tonality under a unified visual style.
|
|
5705
|
+
* - Cinematic camera movement capabilities: Built-in advanced camera movement modes such as bullet time, Grammy slow motion, 360° surround rotation, and full-body model rotation display, achieving deterministic output through precise segmented control.
|
|
5479
5706
|
* - Model library and customization services: Built-in multi-type model resource library, supporting merchants to specify fixed models, customize exclusive model images, and bind brand visual personas for a long time to ensure brand visual consistency.
|
|
5480
|
-
* -
|
|
5707
|
+
* - Adaptation to high-frequency e-commerce content scenarios: Supports multiple resolutions such as 720P and 1080P and multiple aspect ratios such as 3:4 and 9:16, adapting to product detail page video slots, product homepage video displays, waterfall or information feed placements, event venue materials, and new product and promotional video generation.
|
|
5481
5708
|
*
|
|
5482
5709
|
* @param request - VideoGenerationRequest
|
|
5483
5710
|
* @returns VideoGenerationResponse
|
|
@@ -5492,23 +5719,23 @@ class Client extends openapi_core_1.default {
|
|
|
5492
5719
|
* @remarks
|
|
5493
5720
|
* ## Product introduction
|
|
5494
5721
|
* The video translation API translates the speech and on-screen text in a video into a target language. You can select speech subtitle translation, on-screen text translation, or both. This API applies to scenarios such as multilingual distribution of e-commerce videos, international social media marketing, and global brand advertising.
|
|
5495
|
-
* The API uses an asynchronous invocation mode. After you submit a translation task, you can obtain the TaskId and poll the task status by calling the query API until the task is completed and the result is returned.
|
|
5722
|
+
* The API uses an asynchronous invocation mode. After you submit a translation task, you can obtain the `TaskId` and poll the task status by calling the query API until the task is completed and the result is returned.
|
|
5496
5723
|
* ## Scenarios
|
|
5497
|
-
* * **Cross-border e-commerce video localization**: Translates the speech subtitles and selling point text in product introduction videos into the language of the target market to
|
|
5724
|
+
* * **Cross-border e-commerce video localization**: Translates the speech subtitles and selling point text in product introduction videos into the language of the target market to promote products on overseas platforms.
|
|
5498
5725
|
*
|
|
5499
5726
|
* * **International social media marketing**: Localizes speech subtitles and on-screen text for platforms such as TikTok, Instagram, and YouTube with one click to improve comprehension and conversion rates among overseas users.
|
|
5500
5727
|
*
|
|
5501
|
-
* * **Global brand advertising**: Automatically generates video versions in the languages
|
|
5728
|
+
* * **Global brand advertising**: Automatically generates video versions in the corresponding languages based on the target regions to reduce manual production costs.
|
|
5502
5729
|
*
|
|
5503
5730
|
* * **Training and product instructions**: Translates the speech subtitles and on-screen text in training courses or product demo videos into multiple languages for global teams.
|
|
5504
5731
|
* ## Features
|
|
5505
|
-
* |
|
|
5732
|
+
* | Capability | Identifier | Description |
|
|
5506
5733
|
* | --- | --- | --- |
|
|
5507
|
-
* | Speech subtitle translation | speech | Translates the video speech into target language dubbing. You can choose to
|
|
5508
|
-
* | On-screen text translation | visual | Translates the source language text in the video frames, or only erases the on-screen text without translation. You can choose to erase brand logos. |
|
|
5509
|
-
* Speech subtitle translation (speech) generates target language dubbing based on the spoken content in the video, processes the subtitles
|
|
5510
|
-
* On-screen text translation (visual) targets recognizable original video language text in the video frames, such as titles, product selling points, and descriptive text. It can be used independently without selecting speech subtitle translation first, and does not require the text to appear in the speech. Original subtitles within the processing scope may also be processed.
|
|
5511
|
-
* When both speech and visual are selected, the speech subtitle translation is completed and the video is synthesized first
|
|
5734
|
+
* | Speech subtitle translation | `speech` | Translates the video speech into target language dubbing. You can choose to erase the original subtitles only or replace them with translated subtitles. By default, the original subtitles are not processed and no new subtitles are added. |
|
|
5735
|
+
* | On-screen text translation | `visual` | Translates the source language text in the video frames, or only erases the on-screen text without translation. You can choose to erase brand logos. |
|
|
5736
|
+
* Speech subtitle translation (speech) generates target language dubbing based on the spoken content in the video, processes the subtitles based on the original subtitle processing option, and synthesizes the final video. Original subtitles refer to the existing subtitles in the video frames that correspond to the spoken content. Replacing with translated subtitles erases the original subtitles first and then adds subtitles synchronized with the translated speech.
|
|
5737
|
+
* On-screen text translation (visual) targets the recognizable original video language text in the video frames, such as titles, product selling points, and descriptive text. It can be used independently without selecting speech subtitle translation first, and does not require the text to appear in the speech. Original subtitles within the processing scope may also be processed.
|
|
5738
|
+
* When both speech and visual are selected, the speech subtitle translation is completed and the video is synthesized first, and then the on-screen text processing is performed on this video. The translated dubbing is retained, and a final video is output. The order of capabilities in the array does not affect the execution order. Therefore, not processing original subtitles only means that no subtitles are erased or added during the speech subtitle phase, and subsequent on-screen text processing may still change the original subtitles. Subtitles that have been replaced with translated subtitles are retained during subsequent on-screen text processing.
|
|
5512
5739
|
*
|
|
5513
5740
|
* @param tmpReq - VideoTranslationRequest
|
|
5514
5741
|
* @param runtime - runtime options for this request RuntimeOptions
|
|
@@ -5570,23 +5797,23 @@ class Client extends openapi_core_1.default {
|
|
|
5570
5797
|
* @remarks
|
|
5571
5798
|
* ## Product introduction
|
|
5572
5799
|
* The video translation API translates the speech and on-screen text in a video into a target language. You can select speech subtitle translation, on-screen text translation, or both. This API applies to scenarios such as multilingual distribution of e-commerce videos, international social media marketing, and global brand advertising.
|
|
5573
|
-
* The API uses an asynchronous invocation mode. After you submit a translation task, you can obtain the TaskId and poll the task status by calling the query API until the task is completed and the result is returned.
|
|
5800
|
+
* The API uses an asynchronous invocation mode. After you submit a translation task, you can obtain the `TaskId` and poll the task status by calling the query API until the task is completed and the result is returned.
|
|
5574
5801
|
* ## Scenarios
|
|
5575
|
-
* * **Cross-border e-commerce video localization**: Translates the speech subtitles and selling point text in product introduction videos into the language of the target market to
|
|
5802
|
+
* * **Cross-border e-commerce video localization**: Translates the speech subtitles and selling point text in product introduction videos into the language of the target market to promote products on overseas platforms.
|
|
5576
5803
|
*
|
|
5577
5804
|
* * **International social media marketing**: Localizes speech subtitles and on-screen text for platforms such as TikTok, Instagram, and YouTube with one click to improve comprehension and conversion rates among overseas users.
|
|
5578
5805
|
*
|
|
5579
|
-
* * **Global brand advertising**: Automatically generates video versions in the languages
|
|
5806
|
+
* * **Global brand advertising**: Automatically generates video versions in the corresponding languages based on the target regions to reduce manual production costs.
|
|
5580
5807
|
*
|
|
5581
5808
|
* * **Training and product instructions**: Translates the speech subtitles and on-screen text in training courses or product demo videos into multiple languages for global teams.
|
|
5582
5809
|
* ## Features
|
|
5583
|
-
* |
|
|
5810
|
+
* | Capability | Identifier | Description |
|
|
5584
5811
|
* | --- | --- | --- |
|
|
5585
|
-
* | Speech subtitle translation | speech | Translates the video speech into target language dubbing. You can choose to
|
|
5586
|
-
* | On-screen text translation | visual | Translates the source language text in the video frames, or only erases the on-screen text without translation. You can choose to erase brand logos. |
|
|
5587
|
-
* Speech subtitle translation (speech) generates target language dubbing based on the spoken content in the video, processes the subtitles
|
|
5588
|
-
* On-screen text translation (visual) targets recognizable original video language text in the video frames, such as titles, product selling points, and descriptive text. It can be used independently without selecting speech subtitle translation first, and does not require the text to appear in the speech. Original subtitles within the processing scope may also be processed.
|
|
5589
|
-
* When both speech and visual are selected, the speech subtitle translation is completed and the video is synthesized first
|
|
5812
|
+
* | Speech subtitle translation | `speech` | Translates the video speech into target language dubbing. You can choose to erase the original subtitles only or replace them with translated subtitles. By default, the original subtitles are not processed and no new subtitles are added. |
|
|
5813
|
+
* | On-screen text translation | `visual` | Translates the source language text in the video frames, or only erases the on-screen text without translation. You can choose to erase brand logos. |
|
|
5814
|
+
* Speech subtitle translation (speech) generates target language dubbing based on the spoken content in the video, processes the subtitles based on the original subtitle processing option, and synthesizes the final video. Original subtitles refer to the existing subtitles in the video frames that correspond to the spoken content. Replacing with translated subtitles erases the original subtitles first and then adds subtitles synchronized with the translated speech.
|
|
5815
|
+
* On-screen text translation (visual) targets the recognizable original video language text in the video frames, such as titles, product selling points, and descriptive text. It can be used independently without selecting speech subtitle translation first, and does not require the text to appear in the speech. Original subtitles within the processing scope may also be processed.
|
|
5816
|
+
* When both speech and visual are selected, the speech subtitle translation is completed and the video is synthesized first, and then the on-screen text processing is performed on this video. The translated dubbing is retained, and a final video is output. The order of capabilities in the array does not affect the execution order. Therefore, not processing original subtitles only means that no subtitles are erased or added during the speech subtitle phase, and subsequent on-screen text processing may still change the original subtitles. Subtitles that have been replaced with translated subtitles are retained during subsequent on-screen text processing.
|
|
5590
5817
|
*
|
|
5591
5818
|
* @param request - VideoTranslationRequest
|
|
5592
5819
|
* @returns VideoTranslationResponse
|
|
@@ -5596,35 +5823,38 @@ class Client extends openapi_core_1.default {
|
|
|
5596
5823
|
return await this.videoTranslationWithOptions(request, runtime);
|
|
5597
5824
|
}
|
|
5598
5825
|
/**
|
|
5599
|
-
* Generates virtual try-on images by using AI algorithms
|
|
5826
|
+
* Generates virtual try-on images by using AI algorithms to replace a model\\"s clothing with a specified garment. The service preserves the model\\"s pose, facial features, and background details while ensuring a realistic and well-fitted appearance. Users only need to provide the URLs of the original model and garment images to obtain the processed images.
|
|
5600
5827
|
*
|
|
5601
5828
|
* @remarks
|
|
5829
|
+
* *Note:**
|
|
5830
|
+
* VirtualTryOn is an early version that supports only single-item try-on and is scheduled to be deprecated soon.
|
|
5831
|
+
* New features such as multi-item outfit matching, model replacement, and pose replacement are now available in [Virtual try-on (AiTryOn)](https://api.aliyun.com/document/Aidge/2026-04-28/AiTryOn). Refer to the new API for integration.
|
|
5602
5832
|
* ## Scenarios
|
|
5603
|
-
* - **Virtual try-on for
|
|
5604
|
-
* Supports various
|
|
5605
|
-
* - **Virtual try-on for
|
|
5606
|
-
* Supports virtual try-on for
|
|
5833
|
+
* - **Virtual try-on for apparel:**
|
|
5834
|
+
* Supports various apparel categories such as tops, bottoms, jumpsuits, and dresses. After users upload a model image and a product image, the service generates natural and realistic try-on effects to intuitively display how the garment looks on the model.
|
|
5835
|
+
* - **Virtual try-on for footwear and headwear:**
|
|
5836
|
+
* Supports virtual try-on for products such as shoes and hats. The service intelligently matches the product position and proportion based on the model\\"s pose and body parts to present a coordinated and realistic wearing effect.
|
|
5607
5837
|
* - **E-commerce product display and marketing:**
|
|
5608
|
-
*
|
|
5838
|
+
* Helps merchants quickly generate model try-on images for different products. This reduces the costs of traditional photography, outfit changes, and post-production, and improves the production efficiency of product images.
|
|
5609
5839
|
* - **Online shopping experience optimization:**
|
|
5610
|
-
* Allows consumers to preview the wearing effect of products through virtual try-on
|
|
5840
|
+
* Allows consumers to preview the wearing effect of products through virtual try-on. This provides a more intuitive understanding of the product style, matching, and overall effect to assist in purchase decisions and enhance the shopping experience.
|
|
5611
5841
|
* ## Features
|
|
5612
|
-
* - **First,
|
|
5613
|
-
* - a. Model image:
|
|
5614
|
-
* - b. Product image:
|
|
5615
|
-
* - **Second,
|
|
5616
|
-
* - a. Specify product type:
|
|
5617
|
-
* - b. Automatic type recognition:
|
|
5618
|
-
* - **Finally, the system combines the model image, product image, and product type to generate a virtual try-on image,
|
|
5619
|
-
* - a. Preserve model details:
|
|
5620
|
-
* - b. Maintain image consistency:
|
|
5621
|
-
* - c. Generate try-on effect:
|
|
5842
|
+
* - **First, users need to input a model image and a product image. The system uses the model image as the try-on subject and applies the garment from the product image to the model:**
|
|
5843
|
+
* - a. Model image: The image of the person for the virtual try-on.
|
|
5844
|
+
* - b. Product image: The image of the garment, shoes, or hat to be worn by the model.
|
|
5845
|
+
* - **Second, users can select the corresponding product type based on the actual product. Supported types include tops, bottoms, jumpsuits, dresses, shoes, and hats:**
|
|
5846
|
+
* - a. Specify product type: Users can directly input the type of the current product. The system will process it according to the specified type.
|
|
5847
|
+
* - b. Automatic type recognition: If users do not input the product type, the system will automatically recognize the product category in the product image and perform the virtual try-on based on the recognition result.
|
|
5848
|
+
* - **Finally, the system combines the model image, product image, and product type to generate a virtual try-on image, displaying the wearing effect of the product on the model:**
|
|
5849
|
+
* - a. Preserve model details: The generation process does not alter the appearance, hairstyle, pose, body shape, or other details of the person in the original model image.
|
|
5850
|
+
* - b. Maintain image consistency: The background, composition, lighting, and overall visual effect of the original model image remain unchanged. Only the wearing area corresponding to the product is processed.
|
|
5851
|
+
* - c. Generate try-on effect: The product is naturally applied to the model, fitting the pose and body structure of the person to generate a coordinated and realistic virtual try-on image.
|
|
5622
5852
|
* ## Examples
|
|
5623
5853
|
* | **Type** | **Model image** | **Garment image** | **Try-on result** |
|
|
5624
5854
|
* | --- | --- | --- | --- |
|
|
5625
|
-
* | Tops |  |  |  |  |
|
|
5626
5856
|
* | Bottoms |  |  |  |
|
|
5627
|
-
* | Shoes |  |  |  |  |
|
|
5628
5858
|
* | Jumpsuits |  |  |  |
|
|
5629
5859
|
* | Dresses |  |  |  |
|
|
5630
5860
|
* | Hats |  |  |  |
|
|
@@ -5665,35 +5895,38 @@ class Client extends openapi_core_1.default {
|
|
|
5665
5895
|
return $dara.cast(await this.callApi(params, req, runtime), new $_model.VirtualTryOnResponse({}));
|
|
5666
5896
|
}
|
|
5667
5897
|
/**
|
|
5668
|
-
* Generates virtual try-on images by using AI algorithms
|
|
5898
|
+
* Generates virtual try-on images by using AI algorithms to replace a model\\"s clothing with a specified garment. The service preserves the model\\"s pose, facial features, and background details while ensuring a realistic and well-fitted appearance. Users only need to provide the URLs of the original model and garment images to obtain the processed images.
|
|
5669
5899
|
*
|
|
5670
5900
|
* @remarks
|
|
5901
|
+
* *Note:**
|
|
5902
|
+
* VirtualTryOn is an early version that supports only single-item try-on and is scheduled to be deprecated soon.
|
|
5903
|
+
* New features such as multi-item outfit matching, model replacement, and pose replacement are now available in [Virtual try-on (AiTryOn)](https://api.aliyun.com/document/Aidge/2026-04-28/AiTryOn). Refer to the new API for integration.
|
|
5671
5904
|
* ## Scenarios
|
|
5672
|
-
* - **Virtual try-on for
|
|
5673
|
-
* Supports various
|
|
5674
|
-
* - **Virtual try-on for
|
|
5675
|
-
* Supports virtual try-on for
|
|
5905
|
+
* - **Virtual try-on for apparel:**
|
|
5906
|
+
* Supports various apparel categories such as tops, bottoms, jumpsuits, and dresses. After users upload a model image and a product image, the service generates natural and realistic try-on effects to intuitively display how the garment looks on the model.
|
|
5907
|
+
* - **Virtual try-on for footwear and headwear:**
|
|
5908
|
+
* Supports virtual try-on for products such as shoes and hats. The service intelligently matches the product position and proportion based on the model\\"s pose and body parts to present a coordinated and realistic wearing effect.
|
|
5676
5909
|
* - **E-commerce product display and marketing:**
|
|
5677
|
-
*
|
|
5910
|
+
* Helps merchants quickly generate model try-on images for different products. This reduces the costs of traditional photography, outfit changes, and post-production, and improves the production efficiency of product images.
|
|
5678
5911
|
* - **Online shopping experience optimization:**
|
|
5679
|
-
* Allows consumers to preview the wearing effect of products through virtual try-on
|
|
5912
|
+
* Allows consumers to preview the wearing effect of products through virtual try-on. This provides a more intuitive understanding of the product style, matching, and overall effect to assist in purchase decisions and enhance the shopping experience.
|
|
5680
5913
|
* ## Features
|
|
5681
|
-
* - **First,
|
|
5682
|
-
* - a. Model image:
|
|
5683
|
-
* - b. Product image:
|
|
5684
|
-
* - **Second,
|
|
5685
|
-
* - a. Specify product type:
|
|
5686
|
-
* - b. Automatic type recognition:
|
|
5687
|
-
* - **Finally, the system combines the model image, product image, and product type to generate a virtual try-on image,
|
|
5688
|
-
* - a. Preserve model details:
|
|
5689
|
-
* - b. Maintain image consistency:
|
|
5690
|
-
* - c. Generate try-on effect:
|
|
5914
|
+
* - **First, users need to input a model image and a product image. The system uses the model image as the try-on subject and applies the garment from the product image to the model:**
|
|
5915
|
+
* - a. Model image: The image of the person for the virtual try-on.
|
|
5916
|
+
* - b. Product image: The image of the garment, shoes, or hat to be worn by the model.
|
|
5917
|
+
* - **Second, users can select the corresponding product type based on the actual product. Supported types include tops, bottoms, jumpsuits, dresses, shoes, and hats:**
|
|
5918
|
+
* - a. Specify product type: Users can directly input the type of the current product. The system will process it according to the specified type.
|
|
5919
|
+
* - b. Automatic type recognition: If users do not input the product type, the system will automatically recognize the product category in the product image and perform the virtual try-on based on the recognition result.
|
|
5920
|
+
* - **Finally, the system combines the model image, product image, and product type to generate a virtual try-on image, displaying the wearing effect of the product on the model:**
|
|
5921
|
+
* - a. Preserve model details: The generation process does not alter the appearance, hairstyle, pose, body shape, or other details of the person in the original model image.
|
|
5922
|
+
* - b. Maintain image consistency: The background, composition, lighting, and overall visual effect of the original model image remain unchanged. Only the wearing area corresponding to the product is processed.
|
|
5923
|
+
* - c. Generate try-on effect: The product is naturally applied to the model, fitting the pose and body structure of the person to generate a coordinated and realistic virtual try-on image.
|
|
5691
5924
|
* ## Examples
|
|
5692
5925
|
* | **Type** | **Model image** | **Garment image** | **Try-on result** |
|
|
5693
5926
|
* | --- | --- | --- | --- |
|
|
5694
|
-
* | Tops |  |  |  |  |
|
|
5695
5928
|
* | Bottoms |  |  |  |
|
|
5696
|
-
* | Shoes |  |  |  |  |
|
|
5697
5930
|
* | Jumpsuits |  |  |  |
|
|
5698
5931
|
* | Dresses |  |  |  |
|
|
5699
5932
|
* | Hats |  |  |  |
|