hoardboard / real-time-deep-face-recognition

using facenet algorithm

real-time-deep-face-recogniton

Real-time face recognition program using Google's facenet.

• youtube video

Inspiration

• OpenFace
• I refer to the facenet repository of davidsandberg.
• also, shanren7 repository was a great help in implementing.

Dependencies

• Tensorflow 1.2.1 - gpu
• Python 3.5
• Same as requirement.txt in davidsandberg repository.

Pre-trained models

• Inception_ResNet_v1 CASIA-WebFace-> 20170511-185253

Face alignment using MTCNN

• You need det1.npy, det2.npy, and det3.npy in the davidsandberg repository.

How to use

• First, we need align face data. So, if you run 'Make_aligndata.py' first, the face data that is aligned in the 'output_dir' folder will be saved.
• Second, we need to create our own classifier with the face data we created.
  (In the case of me, I had a high recognition rate when I made 30 pictures for each person.)
  Your own classifier is a ~.pkl file that loads the previously mentioned pre-trained model ('20170511-185253.pb') and embeds the face for each person.
  All of these can be obtained by running 'Make_classifier.py'.
  
• Finally, we load our own 'my_classifier.pkl' obtained above and then open the sensor and start recognition.
  (Note that, look carefully at the paths of files and folders in all .py)

Result