added new project type OCRTextlineSegmentation

AI4Bharat · Nov 28, 2024 · 2c36d9e · 2c36d9e
1 parent 4256239
commit 2c36d9e
Show file tree

Hide file tree

Showing 3 changed files with 30 additions and 0 deletions.
diff --git a/backend/projects/annotation_registry.py b/backend/projects/annotation_registry.py
@@ -108,6 +108,15 @@
             "type": "labels",
         },
     },
+    "OCRTextlineSegmentation": {
+        "ocr_transcribed_json": {
+            "to_name": "image_url",
+            "from_name": [
+                "annotation_bboxes",
+            ],
+            "type": ["rectangle"],
+        },
+    },
     "OCRTranscription": {
         "ocr_transcribed_json": {
             "to_name": "image_url",

diff --git a/backend/projects/label_studio_jsx_files/ocr/ocr_textline_segmentation.jsx b/backend/projects/label_studio_jsx_files/ocr/ocr_textline_segmentation.jsx
@@ -0,0 +1,5 @@
+<View>
+  <Style>.ant-input { font-size: large; }</Style>
+  <Image name="image_url" value="$image_url"/>
+  <Rectangle name="annotation_bboxes" toName="image_url" strokeWidth="3" className="ignore_assertion"/>
+</View>
diff --git a/backend/projects/project_registry.yaml b/backend/projects/project_registry.yaml
@@ -99,6 +99,22 @@ OCR:
         fields:
           annotations:
             - ocr_transcribed_json
+    OCRTextlineSegmentation:
+      project_mode: "Annotation"
+      label_studio_jsx_file: "ocr_textline_segmentation.jsx"
+      input_dataset:
+        class: OCRDocument
+        fields:
+          - image_url
+        display_fields:
+          - image_url
+        prediction: ocr_prediction_json
+      output_dataset: 
+        class: OCRDocument
+        save_type: in_place
+        fields:
+          annotations:
+            - ocr_transcribed_json
     OCRTranscriptionEditing:
       project_mode: "Annotation"
       label_studio_jsx_file: "ocr/ocr_transcription.jsx"