Update

app.py

@@ -24,33 +24,53 @@ def process_frame(frame):
     segments = results[0].plot()
     return segments, f'FPS: {int(1 // (end - start))}'
 
-def process_inputs(action, mode_selection, frame, uploaded_video):
+def process_webcam(start_button, stop_button, frame):
     global process, model
-    if action == "Start" and not process and mode_selection:
-        if mode_selection == "Webcam" and frame is not None:
-            process = True
-            model = load_yolo_model()
+    if start_button and not process and frame is not None:
+        process = True
+        model = load_yolo_model()
+        return process_frame(frame)
+    elif stop_button and process:
+        process = False
+        return None, ""
+
+def process_video(start_button, stop_button, uploaded_video):
+    global process, model
+    if start_button and not process and uploaded_video is not None:
+        process = True
+        model = load_yolo_model()
+        cap = cv.VideoCapture(uploaded_video.name)
+        ret, frame = cap.read()
+        if ret:
             return process_frame(frame)
-        elif mode_selection == "Video" and uploaded_video is not None:
-            process = True
-            model = load_yolo_model()
-            return process_frame(uploaded_video)
-    elif action == "Stop" and process:
+    elif stop_button and process:
         process = False
+        return None, ""
 
-action = gr.inputs.Dropdown(choices=["Start", "Stop"], label="Action")
-mode_selection = gr.inputs.Radio(["Webcam", "Video"], label="Mode Selection")
+start_button = gr.inputs.Checkbox(label="Start")
+stop_button = gr.inputs.Checkbox(label="Stop")
 frame = gr.inputs.Video(label="Webcam Feed")
 uploaded_video = gr.inputs.Video(label="Upload Video")
 
-iface = gr.Interface(
-    fn=process_inputs,
-    inputs=[action, mode_selection, frame, uploaded_video],
+webcam_interface = gr.Interface(
+    fn=process_webcam,
+    inputs=[start_button, stop_button, frame],
+    outputs=[gr.components.Image(), gr.components.Textbox()],
+    live=True,
+    title="YOLO Image Segmentation (Webcam)",
+    description="This application uses the YOLO model to perform image segmentation on a webcam feed. Check 'Start' to begin. Check 'Stop' to end the process.",
+    theme="huggingface"
+)
+
+video_interface = gr.Interface(
+    fn=process_video,
+    inputs=[start_button, stop_button, uploaded_video],
     outputs=[gr.components.Image(), gr.components.Textbox()],
     live=True,
-    title="YOLO Image Segmentation",
-    description="This application uses the YOLO model to perform image segmentation on a webcam feed or an uploaded video. Select 'Webcam' or 'Video', upload a video (if applicable), and select 'Start' to begin. Select 'Stop' to end the process.",
+    title="YOLO Image Segmentation (Video)",
+    description="This application uses the YOLO model to perform image segmentation on an uploaded video. Upload a video and check 'Start' to begin. Check 'Stop' to end the process.",
     theme="huggingface"
 )
 
-iface.launch()
+webcam_interface.launch()
+video_interface.launch()