Update app.py

app.py

@@ -254,11 +254,16 @@ def get_detection_data(image_data):
         img = np.array(img)
         img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
 
-        # Apply model size and confidence threshold
-        print(f"Applying default settings - Model Size: {model_size}, Confidence Threshold: {confidence_threshold}")
-        model_selector(model_size)  # Update models based on input
-        update_confidence_threshold(confidence_threshold * 100)  # Scale to 0-100 for consistency
-        print(f"Received new settings - Model Size: {model_size}, Confidence Threshold: {confidence_threshold}")
+        # Log received values for reference (settings applied by prior API calls)
+        print(f"Received settings from payload (reference) - Model Size: {model_size}, Confidence Threshold: {confidence_threshold}")
+        # Verify actual models and threshold used
+        expected_yolo = "yolov8s-seg" if model_size == "Small - Better performance and less accuracy" else \
+                       "yolov8m-seg" if model_size == "Medium - Balanced performance and accuracy" else "yolov8l-seg"
+        expected_midas = "midas_v21_small_256" if model_size == "Small - Better performance and less accuracy" else \
+                        "dpt_hybrid_384" if model_size == "Medium - Balanced performance and accuracy" else "dpt_large_384"
+        print(f"Expected YOLO model: {expected_yolo}, Actual YOLO model: {img_seg.model_type}")
+        print(f"Expected MiDaS model: {expected_midas}, Actual MiDaS model: {depth_estimator.model_type}")
+        print(f"Actual confidence threshold: {img_seg.confidence_threshold}")
 
         #image = utils.resize(img)
         resized_image = utils.resize(img) #depth requires resizing