Update app.py

app.py

@@ -236,12 +236,22 @@ def prod_function(model_pretrained, prod_ds):
     return prod_pred
 
 # Function to get landmarked video
-def tensor_to_video(video_tensor, fps=30):
+def tensor_to_video(video_tensor, fps=30, output_path="output.mp4"):
     video_numpy = video_tensor.permute(0, 2, 3, 1).cpu().numpy()
     # Normalize values to [0, 255] if necessary
     if video_numpy.max() <= 1.0:
-        video_numpy = (video_numpy * 255).astype(np.uint8) 
-    return video_numpy
+        video_numpy = (video_numpy * 255).astype(np.uint8)
+    
+    width = video_numpy.shape[2]
+    height = video_numpy.shape[3]
+    fourcc = cv2.VideoWriter_fourcc(*"mp4v")  # MP4 Codec
+    out = cv2.VideoWriter(output_path, fourcc, fps, (width, height))
+    
+    for frame in prod_video:
+        out.write(frame)
+    out.release()
+    
+    return output_path
 
 # Function to list available videos dynamically
 def list_videos():
@@ -257,7 +267,7 @@ def play_video(selected_video):
 def translate_sign_language(gesture):
     # Create Dataset
     prod_ds = dataset_prod_obj.create_dataset(gesture)
-    prod_video = tensor_to_video(prod_ds)
+    prod_video_path = tensor_to_video(prod_ds)
     #prod_video = np.random.randint(0, 255, (32, 225, 225, 3), dtype=np.uint8)
     
     # Run ML Model
@@ -270,11 +280,11 @@ def translate_sign_language(gesture):
     gesture_translation = idx_to_label[predicted_prod_label.cpu().numpy().item()] # Convert to a scalar
 
     # Frame generator for real-time streaming
-    def frame_generator():
-        for frame in prod_video:
-            yield frame # Stream frame-by-frame
+    #def frame_generator():
+    #    for frame in prod_video:
+    #        yield frame # Stream frame-by-frame
             
-    return gesture_translation , frame_generator() #prod_video
+    return gesture_translation , prod_video_path # frame_generator
 
 with gr.Blocks() as demo:
     gr.Markdown("# Indian Sign Language Translation App")
@@ -284,10 +294,10 @@ with gr.Blocks() as demo:
         with gr.Row(height=350, variant="panel"): # equal_height=False, show_progress=True
             with gr.Column(scale=1, variant="panel"):
                 # Add webcam input for sign language video capture
-                video_input = gr.Video(format="mp4", label="Gesture")
+                video_input = gr.Video(sources=["webcam"], format="mp4", label="Gesture")
             with gr.Column(scale=1, variant="panel"):
                 # Display the landmarked video
-                video_output = gr.Video(streaming=True, label="Landmarked Gesture")
+                video_output = gr.Video(interactive=False, autoplay=True, streaming=False, label="Landmarked Gesture")
         with gr.Row(variant="panel"): # equal_height=False, show_progress=True
             with gr.Column(scale=1, variant="panel"):
                 # Submit the Video