Dynamic-Face-Recognition

Deep learning,Face recognition algorithm based on Mobilenet and Mtcnn(LFW 99.1%, Mobilenet-based face recognition model is superior to the accuracy of the original centerloss paper)

Centorloss&Mtcnn Website

https://github.com/ydwen/caffe-face
https://github.com/dlunion/mtcnn

Environmental requirements

1.Caffe(Be based on CUDA&Cudnn)
2.OpenCV 3.x
3.Cmake

Caffe environment variable setting

1.Add

  export PYTHONPATH=/home/pv/ma/caffe-master/python:$PYTHONPATH
  export CAFFE_ROOT=/home/pv/ma/caffe-master

to ~/.bashrc

Instructions for use

1.Please clip and align the face images that need to be compared and put them in the face_id folder
2.

   cd Dynamic-Face-Recognition
   mkdir build
   cd build
   cmake ..
   ./DFR

DEMO