How it Works
The frontend is built using React. Here users can login/register, upload videos and view their uploaded videos.
The backend consists of a Express Node.js server. Moviepy face-detection, opencv and mediapipe are used to blur the uploaded videos. The face detection function uses the face_locations API from face_recognition library, while the background detection function uses the SelfieSegmentation API from mediapipe library. Then, opencv's GaussianBlur blurs the detected areas and moviepy creates a new processed video.
Uploaded videos and the processed (blurred) videos are stored in a private AWS S3 bucket. This bucket should be accessed by a role following minimum permissions best practice.