Face detection + Face blurring
- Detect all the faces in the input video
- Blur all the detected faces and save them
Code Author: Shih-Yao (Mike) Lin
- OpenCV >= 3.2.0.8
- Tensorflow >= 1.4.1
- MTCNN
- Clone this repo
git clone https://github.com/mikeshihyaolin/face_blurring.gitThe directory tree should look like this:
${ROOT}
├── README.md
├── face_blurring.py
├── figs
│ ├── .DS_Store
│ ├── .blur1.png.icloud
│ ├── blur2.jpg
│ └── blur3.jpg
└── mtcnn
├── .DS_Store
├── __init__.py
├── __pycache__
│ ├── __init__.cpython-36.pyc
│ ├── layer_factory.cpython-36.pyc
│ ├── mtcnn.cpython-36.pyc
│ └── network.cpython-36.pyc
├── data
│ └── .mtcnn_weights.npy.icloud
├── exceptions
│ ├── __init__.py
│ └── __pycache__
│ └── __init__.cpython-36.pyc
├── layer_factory.py
├── mtcnn.py
└── network.py
- Install dependencies
pip3 install -r ./requirements.txt
python face_blurring.py --source_video [input video] --output_blur_img_folder [output folder]
For example,
python face_blurring.py --source_video ./example/video.mp4 --output_video ./blur
The face detection process is heavily borrowed from a MTCNN implementation (Link)