oidn-web 0.2.1 → 0.3.0

package/README.md

@@ -1,17 +1,24 @@
 # Open Image Denoise on the Web
 
 This library brings the state-of-the-art AI-based denoising library [Open Image Denoise](https://github.com/RenderKit/oidn) to the web.
+Currently it's only available on the browsers support WebGPU.
 
-Many thanks to Max Liani for his series https://maxliani.wordpress.com/2023/03/17/dnnd-1-a-deep-neural-network-dive/. My work is mostly inspired by it.
+It's used in the [Vector to 3D](https://www.figma.com/community/plugin/1264600219316901594/) Figma plugin for high quality rendering and denoising.
+
+|                                           2000 Samples                                           |                                     3 Samples                                      |                                   3 Samples + Denoised                                   |
+| :----------------------------------------------------------------------------------------------: | :--------------------------------------------------------------------------------: | :--------------------------------------------------------------------------------------: |
+| ![](https://github.com/pissang/oidn-web/blob/main/examples/test/ground-truth.png 'Ground Truth') | ![](https://github.com/pissang/oidn-web/blob/main/examples/test/noisy.png 'Noisy') | ![](https://github.com/pissang/oidn-web/blob/main/examples/test/denoised.png 'Denoised') |
 
 ## How it Works.
 
 It uses [tfjs](https://github.com/tensorflow/tfjs) to build the UNet model used by the OIDN. Then use the model to do prediction with a WebGPU backend from the image data.
 
-Currently it's only available on the browser with WebGPU enabled.
-
 ## How to Use
 
+[Basic Example](https://oidn-web-example.vercel.app/) ([Code](https://github.com/pissang/oidn-web-example/blob/main/src/main.js))
+
+[Use with three-gpu-pathtracer](https://oidn-web-example.vercel.app/three-gpu-pathtracer.html) ([Code](https://github.com/pissang/oidn-web-example/blob/main/src/three-gpu-pathtracer.js))
+
 ### Install
 
 ```shell
@@ -69,7 +76,7 @@ initUNetFromURL('./weights/rt_hdr.tza', undefined, {
 });
 ```
 
-### Using AUX Input
+### Use auxiliary images
 
 ```ts
 import { UNet, initUNetFromURL } from 'oidn-web';
@@ -98,6 +105,8 @@ initUNetFromURL('./weights/rt_hdr_alb_nrm.tza', undefined, {
 
 If you already have a WebGPU path tracer. You can integrate the oidn-web into your pipeline. It supports input/output gpu buffers to avoid the cost of syncing between CPU and GPU.
 
+`hdr` and `aux` are required in the WebGPU pipeline.
+
 ```ts
 initUNetFromURL(
   './weights/rt_hdr_alb_nrm.tza',
@@ -143,3 +152,7 @@ initUNetFromURL('./weights/rt_hdr_alb_nrm_small.tza', ...);
 ```
 
 Other combinations can be found in the [oidn-weights](https://github.com/RenderKit/oidn-weights)
+
+## Credits
+
+Huge thanks to Max Liani for his series: https://maxliani.wordpress.com/2023/03/17/dnnd-1-a-deep-neural-network-dive/. My work is mostly inspired by it.