This is a mxnet implementation of Zhang's work: Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Neural Networks. It's fast and accurate, see link. This implementation of MTCNN should have almost the same output with the original work.

This repository has been tested under the following environment:

• Python 2.7
• Ubuntu 18.04
• Mxnet-cu90 (==1.3.0)

• Use python main.py to test this detection and alignment method.

• You can change ctx to mx.gpu(0) to use GPU for faster detection.

see mtcnn_detector.py for the details about the parameters. this function use dlib's align strategy, which works well on profile images :)

MIT LICENSE

@article{Zhang2016Joint,
  title={Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks},
  author={Zhang, Kaipeng and Zhang, Zhanpeng and Li, Zhifeng and Yu, Qiao},
  journal={IEEE Signal Processing Letters},
  volume={23},
  number={10},
  pages={1499-1503},
  year={2016},
}

The code is adapted based on an intial fork from the mxnet_mtcnn_face_detection repository.