Update app.py

app.py

@@ -18,7 +18,7 @@ processor = Owlv2Processor.from_pretrained(
 )
 
 # ===============================
-# YOUR PREPROCESSING
+# PREPROCESSING
 # ===============================
 def advanced_preprocessing(img_array: np.ndarray,
                            crop_ratio=(0.25, 0.75),
@@ -40,8 +40,8 @@ def advanced_preprocessing(img_array: np.ndarray,
 
     img_stretch = np.zeros_like(img_resized)
     for c in range(3):
-        img_stretch[:,:,c] = cv2.normalize(
-            img_resized[:,:,c], None, 0, 255, cv2.NORM_MINMAX
+        img_stretch[:, :, c] = cv2.normalize(
+            img_resized[:, :, c], None, 0, 255, cv2.NORM_MINMAX
         )
 
     if tile != (1,1):
@@ -51,13 +51,14 @@ def advanced_preprocessing(img_array: np.ndarray,
 
 
 # ===============================
-# MAIN FUNCTION
+# MAIN FUNCTION (ONLY BOXES)
 # ===============================
 def query_image(img, text_queries, score_threshold):
 
-    # preprocess
+    # preprocess image
     img = advanced_preprocessing(img)
 
+    # still needed internally (model requirement)
     text_queries = [q.strip() for q in text_queries.split(",")]
 
     inputs = processor(
@@ -69,7 +70,6 @@ def query_image(img, text_queries, score_threshold):
     with torch.no_grad():
         outputs = model(**inputs)
 
-    # IMPORTANT FIX
     target_sizes = torch.tensor([img.shape[:2]])
 
     results = processor.post_process_grounded_object_detection(
@@ -79,37 +79,22 @@ def query_image(img, text_queries, score_threshold):
 
     boxes = results["boxes"]
     scores = results["scores"]
-    labels = results["labels"]
 
     output_data = []
 
-    # Draw boxes
-    for box, score, label in zip(boxes, scores, labels):
+    # ONLY bounding boxes
+    for box, score in zip(boxes, scores):
+
         if score < score_threshold:
             continue
 
         x1, y1, x2, y2 = map(int, box.tolist())
-        class_name = text_queries[label.item()]
-        conf = float(score)
 
-        # Save structured output
-        output_data.append({
-            "box": [x1, y1, x2, y2],
-            "label": class_name,
-            "score": round(conf, 3)
-        })
+        # store only coordinates
+        output_data.append([x1, y1, x2, y2])
 
-        # Draw on image
+        # draw rectangle ONLY (no labels)
         cv2.rectangle(img, (x1, y1), (x2, y2), (0,255,0), 2)
-        cv2.putText(
-            img,
-            f"{class_name} {conf:.2f}",
-            (x1, y1-5),
-            cv2.FONT_HERSHEY_SIMPLEX,
-            0.5,
-            (0,255,0),
-            2
-        )
 
     return img, output_data
 
@@ -125,10 +110,10 @@ demo = gr.Interface(
         gr.Slider(0, 1, value=0.2)
     ],
     outputs=[
-        gr.Image(label="Result"),
-        gr.JSON(label="Detections")
+        gr.Image(label="Bounding Boxes"),
+        gr.JSON(label="Coordinates Only")
     ],
-    title="Correct Bounding Box Detection (OWLv2)"
+    title="Bounding Box Coordinates Only (OWLv2)"
 )
 
 demo.launch()