The Video Auto-clipper App is a Streamlit-based application designed for automating the process of locating and clipping specific characters or actions in video files. Developed as part of the NYU CSCI-GA 2565 Machine Learning course, this app utilizes advanced machine learning techniques for face and action recognition to streamline video editing.
Try the live app: Video Auto-clipper or watch the demo video.
- username: admin
- pwd: admin
- Lubin Sun
- Cedric Ni
- Zhiheng Wang
- Helen Zhou
- Face Recognition: Automatically identify and extract video segments featuring a specific character.
- Time-saving: Provides timestamps for each character's appearance to ease the editing process.
- (future)Action Recognition: Locate and clip scenes based on particular actions within the video.
- Upload your video file to the app.
- Click 'Process Video'.
- Let the app process the video and provide you with the clips.
The app uses streamlit, enabling it to analyze video content efficiently. Face and action recognition are powered by MTCNN.
To run the app locally, follow these steps:
git clone [repository-link]
cd [repository-name]
pip install -r requirements.txt # and all other libraries you may be missing
streamlit run app.py