Fixed Sound for Videos

app.py

@@ -24,7 +24,6 @@ if mode == "Image":
 
         st.image(cv2.cvtColor(result, cv2.COLOR_BGR2RGB), caption="Blurred Image")
 
-        # Save to temp file and create download button
         _, temp_path = tempfile.mkstemp(suffix=".jpg")
         cv2.imwrite(temp_path, result)
         with open(temp_path, "rb") as file:
@@ -49,10 +48,10 @@ elif mode == "Video":
 
             with st.spinner("Processing video..."):
                 try:
-                    blur_faces_video(input_path, output_path, update_callback=update_progress)
+                    final_output = blur_faces_video(input_path, output_path, update_callback=update_progress)
                     st.success("✅ Processing complete!")
-                    st.video(output_path)
-                    with open(output_path, "rb") as file:
+                    st.video(final_output)
+                    with open(final_output, "rb") as file:
                         st.download_button("📥 Download Blurred Video", file, file_name="blurred_video.mp4", mime="video/mp4")
                 except Exception as e:
                     st.error(f"Something went wrong: {e}")

requirements.txt

@@ -3,3 +3,5 @@ opencv-python
 ultralytics
 numpy
 Pillow
+moviepy==1.0.3
+imageio-ffmpeg

utils.py

@@ -1,7 +1,7 @@
 import cv2
 from ultralytics import YOLO
+from moviepy.editor import VideoFileClip
 
-# Load your preferred face detection model
 model = YOLO("yolov8-face-hf.pt")
 
 # ------------------ IMAGE FUNCTION ------------------ #
@@ -36,7 +36,8 @@ def blur_faces_video(input_path, output_path, update_callback=None):
     height = height if height % 2 == 0 else height - 1
 
     fourcc = cv2.VideoWriter_fourcc(*'mp4v')
-    out = cv2.VideoWriter(output_path, fourcc, fps, (width, height))
+    temp_video_path = "blurred_temp.mp4"
+    out = cv2.VideoWriter(temp_video_path, fourcc, fps, (width, height))
 
     frame_num = 0
     while True:
@@ -50,7 +51,6 @@ def blur_faces_video(input_path, output_path, update_callback=None):
 
         for box in results.boxes.xyxy:
             x1, y1, x2, y2 = map(int, box)
-
             x1 = max(0, min(w, x1))
             x2 = max(0, min(w, x2))
             y1 = max(0, min(h, y1))
@@ -58,7 +58,6 @@ def blur_faces_video(input_path, output_path, update_callback=None):
 
             face = frame[y1:y2, x1:x2]
             if face.size > 0:
-                # Stronger blur
                 blur_size = max(55, ((x2 - x1) // 2) | 1)
                 if blur_size % 2 == 0:
                     blur_size += 1
@@ -72,4 +71,13 @@ def blur_faces_video(input_path, output_path, update_callback=None):
             update_callback(frame_num / total_frames)
 
     cap.release()
-    out.release()
+    out.release()
+
+    try:
+        original = VideoFileClip(input_path)
+        processed = VideoFileClip(temp_video_path).set_audio(original.audio)
+        processed.write_videofile(output_path, codec="libx264", audio_codec="aac", verbose=False, logger=None)
+        return output_path
+    except Exception as e:
+        print("Audio merging failed:", e)
+        return temp_video_path