Update

app.py

@@ -24,53 +24,35 @@ def process_frame(frame):
     segments = results[0].plot()
     return segments, f'FPS: {int(1 // (end - start))}'
 
-def process_webcam(start_button, stop_button, frame):
+def process_inputs(start_button, stop_button, mode_selection, frame, uploaded_video):
     global process, model
-    if start_button and not process and frame is not None:
+    if start_button and not process:
         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:
+        if mode_selection == "Webcam" and frame is not None:
             return process_frame(frame)
+        elif mode_selection == "Video" and uploaded_video is not None:
+            cap = cv.VideoCapture(uploaded_video.name)
+            ret, frame = cap.read()
+            if ret:
+                return process_frame(frame)
     elif stop_button and process:
         process = False
         return None, ""
 
 start_button = gr.components.Checkbox(label="Start")
 stop_button = gr.components.Checkbox(label="Stop")
+mode_selection = gr.components.Radio(["Webcam", "Video"], label="Mode Selection")
 frame = gr.components.Video(label="Webcam Feed")
 uploaded_video = gr.components.Video(label="Upload 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.",
-)
-
-video_interface = gr.Interface(
-    fn=process_video,
-    inputs=[start_button, stop_button, uploaded_video],
+iface = gr.Interface(
+    fn=process_inputs,
+    inputs=[start_button, stop_button, mode_selection, frame, uploaded_video],
     outputs=[gr.components.Image(), gr.components.Textbox()],
     live=True,
-    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.",
+    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 check 'Start' to begin. Check 'Stop' to end the process.",
 )
 
-gr.TabbedInterface(
-    [webcam_interface, video_interface],
-    tab_names=['Webcam inference', 'Video inference']
-).launch(share=True)
+iface.launch(share=True)