deepdoctection 0.39.7__py3-none-any.whl → 0.41.0__py3-none-any.whl

deepdoctection/pipe/common.py

@@ -22,6 +22,7 @@ from __future__ import annotations
 
 import os
 from copy import deepcopy
+from dataclasses import dataclass, field
 from typing import Literal, Mapping, Optional, Sequence, Union
 
 import numpy as np
@@ -49,24 +50,31 @@ class ImageCroppingService(PipelineComponent):
     generally not stored.
     """
 
-    def __init__(self, category_names: Union[TypeOrStr, Sequence[TypeOrStr]]):
+    def __init__(
+        self,
+        category_names: Optional[Union[TypeOrStr, Sequence[TypeOrStr]]] = None,
+        service_ids: Optional[Sequence[str]] = None,
+    ) -> None:
         """
         :param category_names: A single name or a list of category names to crop
         """
-
-        self.category_names = (
-            (category_names,)
-            if isinstance(category_names, str)
-            else tuple(get_type(category_name) for category_name in category_names)
-        )
+        if category_names is None:
+            self.category_names = None
+        else:
+            self.category_names = (
+                (category_names,)
+                if isinstance(category_names, str)
+                else tuple(get_type(category_name) for category_name in category_names)
+            )
+        self.service_ids = service_ids
         super().__init__("image_crop")
 
     def serve(self, dp: Image) -> None:
-        for ann in dp.get_annotation(category_names=self.category_names):
+        for ann in dp.get_annotation(category_names=self.category_names, service_ids=self.service_ids):
             dp.image_ann_to_image(ann.annotation_id, crop_image=True)
 
     def clone(self) -> ImageCroppingService:
-        return self.__class__(self.category_names)
+        return self.__class__(self.category_names, self.service_ids)
 
     def get_meta_annotation(self) -> MetaAnnotation:
         return MetaAnnotation(image_annotations=(), sub_categories={}, relationships={}, summaries=())
@@ -124,8 +132,10 @@ class IntersectionMatcher:
     def match(
         self,
         dp: Image,
-        parent_categories: Union[TypeOrStr, Sequence[TypeOrStr]],
-        child_categories: Union[TypeOrStr, Sequence[TypeOrStr]],
+        parent_categories: Optional[Union[TypeOrStr, Sequence[TypeOrStr]]] = None,
+        child_categories: Optional[Union[TypeOrStr, Sequence[TypeOrStr]]] = None,
+        parent_ann_service_ids: Optional[Union[str, Sequence[str]]] = None,
+        child_ann_service_ids: Optional[Union[str, Sequence[str]]] = None,
     ) -> list[tuple[str, str]]:
         """
         The matching algorithm
@@ -133,6 +143,10 @@ class IntersectionMatcher:
         :param dp: datapoint image
         :param parent_categories: list of categories to be used a for parent class. Will generate a child-relationship
         :param child_categories: list of categories to be used for a child class.
+        :param parent_ann_service_ids: Additional filter condition. If some ids are selected, it will ignore all other
+                                        parent candidates which are not in the list.
+        :param child_ann_service_ids: Additional filter condition. If some ids are selected, it will ignore all other
+                                        children candidates which are not in the list.
 
         :return: A list of tuples with parent and child annotation ids
         """
@@ -144,6 +158,8 @@ class IntersectionMatcher:
             threshold=self.threshold,
             use_weighted_intersections=self.use_weighted_intersections,
             max_parent_only=self.max_parent_only,
+            parent_ann_service_ids=parent_ann_service_ids,
+            child_ann_service_ids=child_ann_service_ids,
         )
 
         matched_child_anns = np.take(child_anns, child_index)  # type: ignore
@@ -174,8 +190,10 @@ class NeighbourMatcher:
     def match(
         self,
         dp: Image,
-        parent_categories: Union[TypeOrStr, Sequence[TypeOrStr]],
-        child_categories: Union[TypeOrStr, Sequence[TypeOrStr]],
+        parent_categories: Optional[Union[TypeOrStr, Sequence[TypeOrStr]]] = None,
+        child_categories: Optional[Union[TypeOrStr, Sequence[TypeOrStr]]] = None,
+        parent_ann_service_ids: Optional[Union[str, Sequence[str]]] = None,
+        child_ann_service_ids: Optional[Union[str, Sequence[str]]] = None,
     ) -> list[tuple[str, str]]:
         """
         The matching algorithm
@@ -183,16 +201,54 @@ class NeighbourMatcher:
         :param dp: datapoint image
         :param parent_categories: list of categories to be used a for parent class. Will generate a child-relationship
         :param child_categories: list of categories to be used for a child class.
+        :param parent_ann_service_ids: Additional filter condition. If some ids are selected, it will ignore all other
+                                        parent candidates which are not in the list.
+        :param child_ann_service_ids: Additional filter condition. If some ids are selected, it will ignore all other
+                                        children candidates which are not in the list.
 
         :return: A list of tuples with parent and child annotation ids
         """
 
         return [
             (pair[0].annotation_id, pair[1].annotation_id)
-            for pair in match_anns_by_distance(dp, parent_categories, child_categories)
+            for pair in match_anns_by_distance(
+                dp,
+                parent_ann_category_names=parent_categories,
+                child_ann_category_names=child_categories,
+                parent_ann_service_ids=parent_ann_service_ids,
+                child_ann_service_ids=child_ann_service_ids,
+            )
         ]
 
 
+@dataclass
+class FamilyCompound:
+    """
+    A family compound is a set of parent and child categories that are related by a relationship key. The parent
+    categories will receive a relationship to the child categories.
+    """
+
+    relationship_key: Relationships
+    parent_categories: Optional[Union[ObjectTypes, Sequence[ObjectTypes]]] = field(default=None)
+    child_categories: Optional[Union[ObjectTypes, Sequence[ObjectTypes]]] = field(default=None)
+    parent_ann_service_ids: Optional[Union[str, Sequence[str]]] = field(default=None)
+    child_ann_service_ids: Optional[Union[str, Sequence[str]]] = field(default=None)
+
+    def __post_init__(self) -> None:
+        if isinstance(self.parent_categories, str):
+            self.parent_categories = (get_type(self.parent_categories),)
+        elif self.parent_categories is not None:
+            self.parent_categories = tuple(get_type(parent) for parent in self.parent_categories)
+        if isinstance(self.child_categories, str):
+            self.child_categories = (get_type(self.child_categories),)
+        elif self.child_categories is not None:
+            self.child_categories = tuple(get_type(child) for child in self.child_categories)
+        if isinstance(self.parent_ann_service_ids, str):
+            self.parent_ann_service_ids = (self.parent_ann_service_ids,)
+        if isinstance(self.child_ann_service_ids, str):
+            self.child_ann_service_ids = (self.child_ann_service_ids,)
+
+
 @pipeline_component_registry.register("MatchingService")
 class MatchingService(PipelineComponent):
     """
@@ -202,28 +258,15 @@ class MatchingService(PipelineComponent):
 
     def __init__(
         self,
-        parent_categories: Union[TypeOrStr, Sequence[TypeOrStr]],
-        child_categories: Union[TypeOrStr, Sequence[TypeOrStr]],
+        family_compounds: Sequence[FamilyCompound],
         matcher: Union[IntersectionMatcher, NeighbourMatcher],
-        relationship_key: Relationships,
     ) -> None:
         """
-        :param parent_categories: list of categories to be used a for parent class. Will generate a child-relationship
-        :param child_categories: list of categories to be used for a child class.
-
+        :param family_compounds: A list of FamilyCompounds
+        :param matcher: A matcher object
         """
-        self.parent_categories = (
-            (get_type(parent_categories),)
-            if isinstance(parent_categories, str)
-            else tuple(get_type(category_name) for category_name in parent_categories)
-        )
-        self.child_categories = (
-            (get_type(child_categories),)
-            if isinstance(child_categories, str)
-            else (tuple(get_type(category_name) for category_name in child_categories))
-        )
+        self.family_compounds = family_compounds
         self.matcher = matcher
-        self.relationship_key = relationship_key
         super().__init__("matching")
 
     def serve(self, dp: Image) -> None:
@@ -233,20 +276,31 @@ class MatchingService(PipelineComponent):
 
         :param dp: datapoint image
         """
-
-        matched_pairs = self.matcher.match(dp, self.parent_categories, self.child_categories)
-
-        for pair in matched_pairs:
-            self.dp_manager.set_relationship_annotation(self.relationship_key, pair[0], pair[1])
+        for family_compound in self.family_compounds:
+            matched_pairs = self.matcher.match(
+                dp,
+                parent_categories=family_compound.parent_categories,
+                child_categories=family_compound.child_categories,
+                parent_ann_service_ids=family_compound.parent_ann_service_ids,
+                child_ann_service_ids=family_compound.child_ann_service_ids,
+            )
+
+            for pair in matched_pairs:
+                self.dp_manager.set_relationship_annotation(family_compound.relationship_key, pair[0], pair[1])
 
     def clone(self) -> PipelineComponent:
-        return self.__class__(self.parent_categories, self.child_categories, self.matcher, self.relationship_key)
+        return self.__class__(self.family_compounds, self.matcher)
 
     def get_meta_annotation(self) -> MetaAnnotation:
+        relationships: dict[ObjectTypes, set[ObjectTypes]] = {}
+        for family_compound in self.family_compounds:
+            if family_compound.parent_categories is not None:
+                for parent_category in family_compound.parent_categories:
+                    relationships[parent_category] = {family_compound.relationship_key}  # type: ignore
         return MetaAnnotation(
             image_annotations=(),
             sub_categories={},
-            relationships={parent: {Relationships.CHILD} for parent in self.parent_categories},
+            relationships=relationships,
             summaries=(),
         )
 

deepdoctection/pipe/layout.py

@@ -20,18 +20,41 @@ Module for layout pipeline component
 """
 from __future__ import annotations
 
-from typing import Optional
+from typing import Optional, Sequence, Union
 
 import numpy as np
 
 from ..datapoint.image import Image
 from ..extern.base import ObjectDetector, PdfMiner
+from ..mapper.misc import curry
 from ..utils.error import ImageError
+from ..utils.settings import ObjectTypes
 from ..utils.transform import PadTransform
 from .base import MetaAnnotation, PipelineComponent
 from .registry import pipeline_component_registry
 
 
+@curry
+def skip_if_category_or_service_extracted(
+    dp: Image,
+    category_names: Optional[Union[str, Sequence[ObjectTypes]]] = None,
+    service_ids: Optional[Union[str, Sequence[str]]] = None,
+) -> bool:
+    """
+    Skip the processing of the pipeline component if the category or service is already extracted.
+
+    **Example**
+
+        detector = # some detector
+        item_component = ImageLayoutService(detector)
+        item_component.set_inbound_filter(skip_if_category_or_service_extracted(detector.get_categories(as_dict=False)))
+    """
+
+    if dp.get_annotation(category_names=category_names, service_ids=service_ids):
+        return True
+    return False
+
+
 @pipeline_component_registry.register("ImageLayoutService")
 class ImageLayoutService(PipelineComponent):
     """
@@ -45,7 +68,7 @@ class ImageLayoutService(PipelineComponent):
 
     **Example**
 
-            d_items = TPFrcnnDetector(item_config_path, item_weights_path, {"1": "ROW", "2": "COLUMNS"})
+            d_items = TPFrcnnDetector(item_config_path, item_weights_path, {1: 'row', 2: 'column'})
             item_component = ImageLayoutService(d_items)
     """
 
@@ -55,7 +78,6 @@ class ImageLayoutService(PipelineComponent):
         to_image: bool = False,
         crop_image: bool = False,
         padder: Optional[PadTransform] = None,
-        skip_if_layout_extracted: bool = False,
     ):
         """
         :param layout_detector: object detector
@@ -65,23 +87,14 @@ class ImageLayoutService(PipelineComponent):
                            to its bounding box and populate the resulting sub image to
                            `ImageAnnotation.image.image`.
         :param padder: If not `None`, will apply the padder to the image before prediction and inverse apply the padder
-        :param skip_if_layout_extracted: When `True` will check, if there are already `ImageAnnotation` of a category
-                                         available that will be predicted by the `layout_detector`. If yes, will skip
-                                         the prediction process.
         """
         self.to_image = to_image
         self.crop_image = crop_image
         self.padder = padder
-        self.skip_if_layout_extracted = skip_if_layout_extracted
         self.predictor = layout_detector
         super().__init__(self._get_name(layout_detector.name), self.predictor.model_id)
 
     def serve(self, dp: Image) -> None:
-        if self.skip_if_layout_extracted:
-            categories = self.predictor.get_category_names()
-            anns = dp.get_annotation(category_names=categories)
-            if anns:
-                return
         if dp.image is None:
             raise ImageError("image cannot be None")
         np_image = dp.image
@@ -117,7 +130,7 @@ class ImageLayoutService(PipelineComponent):
             padder_clone = self.padder.clone()
         if not isinstance(predictor, ObjectDetector):
             raise TypeError(f"predictor must be of type ObjectDetector, but is of type {type(predictor)}")
-        return self.__class__(predictor, self.to_image, self.crop_image, padder_clone, self.skip_if_layout_extracted)
+        return self.__class__(predictor, self.to_image, self.crop_image, padder_clone)
 
     def clear_predictor(self) -> None:
         self.predictor.clear_model()

deepdoctection/pipe/order.py

@@ -347,19 +347,15 @@ class TextLineGenerator:
     a paragraph break threshold. This allows to detect a multi column structure just by observing sub lines.
     """
 
-    def __init__(
-        self, make_sub_lines: bool, line_category_id: Union[int, str], paragraph_break: Optional[float] = None
-    ):
+    def __init__(self, make_sub_lines: bool, paragraph_break: Optional[float] = None):
         """
         :param make_sub_lines: Whether to build sub lines from lines.
-        :param line_category_id: category_id to give a text line
         :param paragraph_break: threshold of two consecutive words. If distance is larger than threshold, two sub-lines
                                 will be built. We use relative coordinates to calculate the distance between two
                                 consecutive words. A reasonable value is 0.035
         """
         if make_sub_lines and paragraph_break is None:
             raise ValueError("You must specify paragraph_break when setting make_sub_lines to True")
-        self.line_category_id = int(line_category_id)
         self.make_sub_lines = make_sub_lines
         self.paragraph_break = paragraph_break
 
@@ -367,7 +363,6 @@ class TextLineGenerator:
         return DetectionResult(
             box=box.to_list(mode="xyxy"),
             class_name=LayoutType.LINE,
-            class_id=self.line_category_id,
             absolute_coords=box.absolute_coords,
             relationships=relationships,
         )
@@ -475,18 +470,14 @@ class TextLineServiceMixin(PipelineComponent, ABC):
     def __init__(
         self,
         name: str,
-        line_category_id: int = 1,
         include_residual_text_container: bool = True,
         paragraph_break: Optional[float] = None,
     ):
         """
-        Initialize the TextLineService with a line_category_id and a TextLineGenerator instance.
+        Initialize the TextLineServiceMixin with a TextLineGenerator instance.
         """
-        self.line_category_id = line_category_id
         self.include_residual_text_container = include_residual_text_container
-        self.text_line_generator = TextLineGenerator(
-            self.include_residual_text_container, self.line_category_id, paragraph_break
-        )
+        self.text_line_generator = TextLineGenerator(self.include_residual_text_container, paragraph_break)
         super().__init__(name)
 
     def _create_lines_for_words(self, word_anns: Sequence[ImageAnnotation]) -> Sequence[ImageAnnotation]:
@@ -523,17 +514,15 @@ class TextLineService(TextLineServiceMixin):
     text lines and the words contained in the text lines. The reading order is not arranged.
     """
 
-    def __init__(self, line_category_id: int = 1, paragraph_break: Optional[float] = None):
+    def __init__(self, paragraph_break: Optional[float] = None):
         """
         Initialize `TextLineService`
 
-        :param line_category_id: category_id to give a text line
         :param paragraph_break: threshold of two consecutive words. If distance is larger than threshold, two sublines
                                 will be built
         """
         super().__init__(
             name="text_line",
-            line_category_id=line_category_id,
             include_residual_text_container=True,
             paragraph_break=paragraph_break,
         )
@@ -542,7 +531,7 @@ class TextLineService(TextLineServiceMixin):
         """
         This method returns a new instance of the class with the same configuration.
         """
-        return self.__class__(self.line_category_id, self.text_line_generator.paragraph_break)
+        return self.__class__(self.text_line_generator.paragraph_break)
 
     def serve(self, dp: Image) -> None:
         text_container_anns = dp.get_annotation(category_names=LayoutType.WORD)
@@ -605,7 +594,6 @@ class TextOrderService(TextLineServiceMixin):
         broken_line_tolerance: float = 0.003,
         height_tolerance: float = 2.0,
         paragraph_break: Optional[float] = 0.035,
-        line_category_id: int = 1,
     ):
         """
         :param text_container: name of an image annotation that has a CHARS sub category. These annotations will be
@@ -647,12 +635,9 @@ class TextOrderService(TextLineServiceMixin):
             self.floating_text_block_categories = self.floating_text_block_categories + (LayoutType.LINE,)
         self.include_residual_text_container = include_residual_text_container
         self.order_generator = OrderGenerator(starting_point_tolerance, broken_line_tolerance, height_tolerance)
-        self.text_line_generator = TextLineGenerator(
-            self.include_residual_text_container, line_category_id, paragraph_break
-        )
+        self.text_line_generator = TextLineGenerator(self.include_residual_text_container, paragraph_break)
         super().__init__(
             name="text_order",
-            line_category_id=line_category_id,
             include_residual_text_container=include_residual_text_container,
             paragraph_break=paragraph_break,
         )
@@ -763,7 +748,6 @@ class TextOrderService(TextLineServiceMixin):
             self.order_generator.broken_line_tolerance,
             self.order_generator.height_tolerance,
             self.text_line_generator.paragraph_break,
-            self.text_line_generator.line_category_id,
         )
 
     def clear_predictor(self) -> None:

deepdoctection/pipe/segment.py

@@ -436,24 +436,24 @@ def segment_table(
     child_ann_ids = table.get_relationship(Relationships.CHILD)
     cell_index_rows, row_index, _, _ = match_anns_by_intersection(
         dp,
-        item_names[0],
-        cell_names,
-        segment_rule,
-        threshold_rows,
-        True,
-        child_ann_ids,
-        child_ann_ids,
+        parent_ann_category_names=item_names[0],
+        child_ann_category_names=cell_names,
+        matching_rule=segment_rule,
+        threshold=threshold_rows,
+        use_weighted_intersections=True,
+        parent_ann_ids=child_ann_ids,
+        child_ann_ids=child_ann_ids,
     )
 
     cell_index_cols, col_index, _, _ = match_anns_by_intersection(
         dp,
-        item_names[1],
-        cell_names,
-        segment_rule,
-        threshold_cols,
-        True,
-        child_ann_ids,
-        child_ann_ids,
+        parent_ann_category_names=item_names[1],
+        child_ann_category_names=cell_names,
+        matching_rule=segment_rule,
+        threshold=threshold_cols,
+        use_weighted_intersections=True,
+        parent_ann_ids=child_ann_ids,
+        child_ann_ids=child_ann_ids,
     )
 
     cells = dp.get_annotation(annotation_ids=child_ann_ids, category_names=cell_names)
@@ -499,7 +499,6 @@ def create_intersection_cells(
     rows: Sequence[ImageAnnotation],
     cols: Sequence[ImageAnnotation],
     table_annotation_id: str,
-    cell_class_id: int,
     sub_item_names: Sequence[ObjectTypes],
 ) -> tuple[Sequence[DetectionResult], Sequence[SegmentationResult]]:
     """
@@ -509,7 +508,6 @@ def create_intersection_cells(
     :param rows: list of rows
     :param cols: list of columns
     :param table_annotation_id: annotation_id of underlying table ImageAnnotation
-    :param cell_class_id: The class_id to a synthetically generated DetectionResult
     :param sub_item_names: ObjectTypes for row-/column number
     :return: Pair of lists of `DetectionResult` and `SegmentationResult`.
     """
@@ -526,7 +524,6 @@ def create_intersection_cells(
             detect_result_cells.append(
                 DetectionResult(
                     box=boxes_cells[idx].to_list(mode="xyxy"),
-                    class_id=cell_class_id,
                     absolute_coords=boxes_cells[idx].absolute_coords,
                     class_name=LayoutType.CELL,
                 )
@@ -574,13 +571,13 @@ def header_cell_to_item_detect_result(
     child_ann_ids = table.get_relationship(Relationships.CHILD)
     item_index, _, items, _ = match_anns_by_intersection(
         dp,
-        item_header_name,
-        item_name,
-        segment_rule,
-        threshold,
-        True,
-        child_ann_ids,
-        child_ann_ids,
+        parent_ann_category_names=item_header_name,
+        child_ann_category_names=item_name,
+        matching_rule=segment_rule,
+        threshold=threshold,
+        use_weighted_intersections=True,
+        parent_ann_ids=child_ann_ids,
+        child_ann_ids=child_ann_ids,
     )
     item_headers = []
     for idx, item in enumerate(items):
@@ -622,24 +619,24 @@ def segment_pubtables(
     child_ann_ids = table.get_relationship(Relationships.CHILD)
     cell_index_rows, row_index, _, _ = match_anns_by_intersection(
         dp,
-        item_names[0],
-        spanning_cell_names,
-        segment_rule,
-        threshold_rows,
-        True,
-        child_ann_ids,
-        child_ann_ids,
+        parent_ann_category_names=item_names[0],
+        child_ann_category_names=spanning_cell_names,
+        matching_rule=segment_rule,
+        threshold=threshold_rows,
+        use_weighted_intersections=True,
+        parent_ann_ids=child_ann_ids,
+        child_ann_ids=child_ann_ids,
     )
 
     cell_index_cols, col_index, _, _ = match_anns_by_intersection(
         dp,
-        item_names[1],
-        spanning_cell_names,
-        segment_rule,
-        threshold_cols,
-        True,
-        child_ann_ids,
-        child_ann_ids,
+        parent_ann_category_names=item_names[1],
+        child_ann_category_names=spanning_cell_names,
+        matching_rule=segment_rule,
+        threshold=threshold_cols,
+        use_weighted_intersections=True,
+        parent_ann_ids=child_ann_ids,
+        child_ann_ids=child_ann_ids,
     )
 
     spanning_cells = dp.get_annotation(annotation_ids=child_ann_ids, category_names=spanning_cell_names)
@@ -976,7 +973,6 @@ class PubtablesSegmentationService(PipelineComponent):
         tile_table_with_items: bool,
         remove_iou_threshold_rows: float,
         remove_iou_threshold_cols: float,
-        cell_class_id: int,
         table_name: TypeOrStr,
         cell_names: Sequence[TypeOrStr],
         spanning_cell_names: Sequence[TypeOrStr],
@@ -997,7 +993,6 @@ class PubtablesSegmentationService(PipelineComponent):
                                       the adjacent row. Will do a similar shifting with columns.
         :param remove_iou_threshold_rows: iou threshold for removing overlapping rows
         :param remove_iou_threshold_cols: iou threshold for removing overlapping columns
-        :param cell_class_id: 'category_id' for cells to be generated from intersected rows and columns
         :param table_name: layout type table
         :param cell_names: layout type of cells
         :param spanning_cell_names: layout type of spanning cells
@@ -1022,7 +1017,6 @@ class PubtablesSegmentationService(PipelineComponent):
         self.spanning_cell_names = [get_type(cell_name) for cell_name in spanning_cell_names]
         self.remove_iou_threshold_rows = remove_iou_threshold_rows
         self.remove_iou_threshold_cols = remove_iou_threshold_cols
-        self.cell_class_id = cell_class_id
         self.cell_to_image = cell_to_image
         self.crop_cell_image = crop_cell_image
         self.item_names = [get_type(item_name) for item_name in item_names]  # row names must be before column name
@@ -1089,7 +1083,7 @@ class PubtablesSegmentationService(PipelineComponent):
             rows = dp.get_annotation(category_names=self.item_names[0], annotation_ids=item_ann_ids)
             columns = dp.get_annotation(category_names=self.item_names[1], annotation_ids=item_ann_ids)
             detect_result_cells, segment_result_cells = create_intersection_cells(
-                rows, columns, table.annotation_id, self.cell_class_id, self.sub_item_names
+                rows, columns, table.annotation_id, self.sub_item_names
             )
             cell_rn_cn_to_ann_id = {}
             for detect_result, segment_result in zip(detect_result_cells, segment_result_cells):
@@ -1228,7 +1222,6 @@ class PubtablesSegmentationService(PipelineComponent):
             self.tile_table,
             self.remove_iou_threshold_rows,
             self.remove_iou_threshold_cols,
-            self.cell_class_id,
             self.table_name,
             self.cell_names,
             self.spanning_cell_names,

deepdoctection/pipe/sub_layout.py

@@ -92,7 +92,6 @@ class DetectResultGenerator:
                         detect_result_list.append(
                             DetectionResult(
                                 box=[0.0, 0.0, float(self.width), float(self.height)],  # type: ignore
-                                class_id=self.categories_name_as_key[category_name],
                                 class_name=category_name,
                                 score=0.0,
                                 absolute_coords=self.absolute_coords,
@@ -154,16 +153,16 @@ class SubImageLayoutService(PipelineComponent):
     **Example**
 
             detect_result_generator = DetectResultGenerator(categories_items)
-            d_items = TPFrcnnDetector(item_config_path, item_weights_path, {"1": LayoutType.row,
-            "2": LayoutType.column})
-            item_component = SubImageLayoutService(d_items, LayoutType.table, {1: 7, 2: 8}, detect_result_generator)
+            d_items = TPFrcnnDetector(item_config_path, item_weights_path, {1: LayoutType.row,
+            2: LayoutType.column})
+            item_component = SubImageLayoutService(d_items, LayoutType.table, detect_result_generator)
     """
 
     def __init__(
         self,
         sub_image_detector: ObjectDetector,
         sub_image_names: Union[str, Sequence[TypeOrStr]],
-        category_id_mapping: Optional[dict[int, int]] = None,
+        service_ids: Optional[Sequence[str]] = None,
         detect_result_generator: Optional[DetectResultGenerator] = None,
         padder: Optional[PadTransform] = None,
     ):
@@ -172,7 +171,8 @@ class SubImageLayoutService(PipelineComponent):
         :param sub_image_names: Category names of ImageAnnotations to be presented to the detector.
                                 Attention: The selected ImageAnnotations must have: attr:`image` and: attr:`image.image`
                                 not None.
-        :param category_id_mapping: Mapping of category IDs. Usually, the category ids start with 1.
+        :param service_ids: List of service ids to be used for filtering the ImageAnnotations. If None, all
+                            ImageAnnotations will be used.
         :param detect_result_generator: 'DetectResultGenerator' instance. 'categories' attribute has to be the same as
                                         the 'categories' attribute of the 'sub_image_detector'. The generator will be
                                         responsible to create 'DetectionResult' for some categories, if they have not
@@ -186,7 +186,7 @@ class SubImageLayoutService(PipelineComponent):
             if isinstance(sub_image_names, str)
             else tuple((get_type(cat) for cat in sub_image_names))
         )
-        self.category_id_mapping = category_id_mapping
+        self.service_ids = service_ids
         self.detect_result_generator = detect_result_generator
         self.padder = padder
         self.predictor = sub_image_detector
@@ -208,7 +208,7 @@ class SubImageLayoutService(PipelineComponent):
         - Optionally invoke the DetectResultGenerator
         - Generate ImageAnnotations and dump to parent image and sub image.
         """
-        sub_image_anns = dp.get_annotation(category_names=self.sub_image_name)
+        sub_image_anns = dp.get_annotation(category_names=self.sub_image_name, service_ids=self.service_ids)
         for sub_image_ann in sub_image_anns:
             np_image = self.prepare_np_image(sub_image_ann)
             detect_result_list = self.predictor.predict(np_image)
@@ -223,11 +223,6 @@ class SubImageLayoutService(PipelineComponent):
                 detect_result_list = self.detect_result_generator.create_detection_result(detect_result_list)
 
             for detect_result in detect_result_list:
-                if self.category_id_mapping:
-                    if detect_result.class_id:
-                        detect_result.class_id = self.category_id_mapping.get(
-                            detect_result.class_id, detect_result.class_id
-                        )
                 self.dp_manager.set_image_annotation(detect_result, sub_image_ann.annotation_id)
 
     def get_meta_annotation(self) -> MetaAnnotation:
@@ -254,7 +249,7 @@ class SubImageLayoutService(PipelineComponent):
         return self.__class__(
             predictor,
             self.sub_image_name,
-            self.category_id_mapping,
+            self.service_ids,
             self.detect_result_generator,
             padder_clone,
         )