Update app.py

app.py

@@ -1,6 +1,10 @@
 import gradio as gr
 import numpy as np
 import time
+import threading
+
+# Global flag to control detection
+detection_active = False
 
 def process_video_frame(frame):
     """
@@ -23,6 +27,34 @@ def process_video_frame(frame):
     
     return status, result, conf_text
 
+def start_detection():
+    """
+    Start the detection process
+    """
+    global detection_active
+    detection_active = True
+    return "Processing live feed..."
+
+def stop_detection():
+    """
+    Stop the detection process
+    """
+    global detection_active
+    detection_active = False
+    return "Detection stopped"
+
+def process_frames(video_feed, status_text, result_text, confidence_text):
+    """
+    Continuously process frames when detection is active
+    """
+    while detection_active:
+        if video_feed is not None:
+            status, result, conf = process_video_frame(video_feed)
+            status_text.update(value=status)
+            result_text.update(value=result)
+            confidence_text.update(value=conf)
+        time.sleep(0.1)  # Adjust the sleep time as needed
+
 with gr.Blocks() as demo:
     gr.Markdown("# Real-Time Face Spoofing Detection")
     
@@ -53,35 +85,23 @@ with gr.Blocks() as demo:
     - Detection results update in real-time
     """)
     
-    # Update interval in seconds (adjust based on your needs)
-    PROCESS_INTERVAL = 0.1
-    
-    # Handle start/stop functionality
-    def start_detection():
-        return gr.update(value="Processing live feed...")
-    
-    def stop_detection():
-        return gr.update(value="Detection stopped")
-    
     # Event handlers
     start_button.click(
         fn=start_detection,
         outputs=status_text
     )
     
-    # Process frames when detection is running
-    video_feed.change(
-        fn=process_video_frame,
-        inputs=[video_feed],
-        outputs=[status_text, result_text, confidence_text],
-        every=PROCESS_INTERVAL
-    )
-    
     stop_button.click(
         fn=stop_detection,
-        outputs=status_text,
-        cancels=[video_feed]
+        outputs=status_text
     )
+    
+    # Start a thread to process frames when detection is active
+    threading.Thread(
+        target=process_frames,
+        args=(video_feed, status_text, result_text, confidence_text),
+        daemon=True
+    ).start()
 
 if __name__ == "__main__":
     demo.launch()