Update app.py

app.py

@@ -16,10 +16,20 @@ def predict(image):
     results_all = model_all(img_array)
     results_np = model_np(img_array)
 
+    # Merge detections into a single image
     # Plot results from the first model (all detections)
+    combined_image = None
     for r in results_all:
         im_array = r.plot()
-    im = Image.fromarray(im_array[..., ::-1])  # RGB PIL image
+        combined_image = im_array
+
+    # Plot results from the second model (number plate detection) on the same image
+    for r in results_np:
+        np_array = r.plot()
+        if combined_image is not None:
+            combined_image = np.maximum(combined_image, np_array)  # Combine both images (max pixel values)
+
+    im = Image.fromarray(combined_image[..., ::-1])  # Convert back to RGB PIL image
 
     # Initialize counters and confidence lists
     class_counts_all = {i: 0 for i in range(len(model_all.names))}
@@ -39,17 +49,16 @@ def predict(image):
         np_count += 1
         np_confidences.append(float(box.conf[0]))
 
-    # Create output string
+    # Combine the counts from both models into one output string
     output = "Detection Results:\n"
-    output += "Model 1 (Helmet, License Plate, Motorcyclist):\n"
     for i in range(len(model_all.names)):
         count = class_counts_all[i]
         avg_conf = np.mean(class_confidences_all[i]) if class_confidences_all[i] else 0
         output += f"{model_all.names[i]}: {count} detections (Avg. Confidence: {avg_conf:.2f})\n"
 
-    output += "\nModel 2 (Number Plate Detection):\n"
+    # Add number plate detection results from the second model
     avg_np_conf = np.mean(np_confidences) if np_confidences else 0
-    output += f"Number Plates: {np_count} detections (Avg. Confidence: {avg_np_conf:.2f})\n"
+    output += f"Number Plates : {np_count} detections (Avg. Confidence: {avg_np_conf:.2f})\n"
 
     return im, output
 
@@ -66,4 +75,4 @@ iface = gr.Interface(
 )
 
 # Launch the interface
-iface.launch(share='True')
+iface.launch(share='True')