Update app.py

app.py

@@ -1,72 +1,72 @@
 import gradio as gr
 import cv2
+import numpy as np
 import mediapipe as mp
-import os
+import logging
+
+# Enable logging for debugging
+logging.basicConfig(level=logging.DEBUG)
 
 # Initialize MediaPipe Hands model
 mp_hands = mp.solutions.hands
 mp_drawing = mp.solutions.drawing_utils
-hands = mp_hands.Hands(min_detection_confidence=0.3, min_tracking_confidence=0.5)  # Lowered threshold
-
-def detect_hands(video):
-    if not video or 'name' not in video:
-        return "Error: No video uploaded."
+hands = mp_hands.Hands(static_image_mode=False, max_num_hands=1, min_detection_confidence=0.5)
 
-    video_path = video['name']
-    print(f"Received video path: {video_path}")
+def detect_hands(video_path):
+    logging.info(f"Received video path: {video_path}")
 
-    if not os.path.exists(video_path):
-        return "Error: Video file not found."
+    # Check if video path is valid
+    if not video_path:
+        logging.error("No video file received.")
+        return "Error: No video file received."
 
+    # Open video file
     cap = cv2.VideoCapture(video_path)
-    frames = []
-    hand_detected = False  # Flag to check if hand detection is working
+    
+    if not cap.isOpened():
+        logging.error("Error opening video file.")
+        return "Error: Unable to open video."
+
+    frame_count = 0
+    detected_frames = 0
 
     while cap.isOpened():
         ret, frame = cap.read()
         if not ret:
-            break  # No more frames to read
+            break  # Exit loop if video ends
         
+        frame_count += 1
+        logging.debug(f"Processing frame {frame_count}")
+
+        # Convert frame to RGB
         frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
         results = hands.process(frame_rgb)
 
-        # Draw landmarks if hands are detected
         if results.multi_hand_landmarks:
-            hand_detected = True  # Set flag to True if a hand is detected
-            print("Hand detected!")
+            detected_frames += 1
+            logging.info(f"Hand detected in frame {frame_count}")
             for hand_landmarks in results.multi_hand_landmarks:
                 mp_drawing.draw_landmarks(frame, hand_landmarks, mp_hands.HAND_CONNECTIONS)
-        else:
-            print("No hand detected in this frame.")
-
-        frames.append(frame)
 
     cap.release()
-    print(f"Total frames processed: {len(frames)}")
-
-    if not hand_detected:
-        return "Error: No hands detected. Try adjusting lighting or hand position."
 
-    if not frames:
-        return "Error: No frames were processed. Try converting your video to .mp4 and re-uploading."
+    # Log summary
+    logging.info(f"Total frames processed: {frame_count}")
+    logging.info(f"Frames with hand detected: {detected_frames}")
 
-    height, width, _ = frames[0].shape
-    output_path = "hand_detected_video.mp4"
-    out = cv2.VideoWriter(output_path, cv2.VideoWriter_fourcc(*'mp4v'), 20, (width, height))
-
-    for frame in frames:
-        out.write(frame)
-    
-    out.release()
-    return output_path
+    if detected_frames > 0:
+        return "Hand detected in the video!"
+    else:
+        return "No hand detected in the video."
 
+# Gradio Interface
 iface = gr.Interface(
     fn=detect_hands,
-    inputs=gr.Video(label="Upload Sign Language Video"),
-    outputs=gr.Video(label="Hand Landmarks Detected"),
-    title="Sign Language Hand Tracking",
-    description="Uploads a video, detects hand landmarks using MediaPipe, and returns the processed video."
+    inputs=gr.Video(),
+    outputs="text",
+    title="Sign Language Translator",
+    description="Upload a video of a sign, and the model will detect hands."
 )
 
 if __name__ == "__main__":
-    iface.launch()
+    iface.launch(debug=True)