Skip to content

Latest commit

 

History

History
28 lines (19 loc) · 973 Bytes

README.md

File metadata and controls

28 lines (19 loc) · 973 Bytes

Face Anti-Spoofing

Several face anti-spoofing models from github. There are eight models to detect spoofing. For real-life projects combination of m6 and m8 gives a high accuracy (more than 99%).

  1. M1: https://github.com/zeusees/HyperFAS
  2. M2: https://github.com/emadeldeen24/face-anti-spoofing
  3. M3: https://github.com/Davidzhangyuanhan/CelebA-Spoof
  4. M4: https://github.com/JinghuiZhou/awesome_face_antispoofing
  5. M5: Simple CNN model. I have trained this model from scratch.
  6. M6: https://github.com/minivision-ai/Silent-Face-Anti-Spoofing
  7. M7: https://github.com/Saiyam26/Face-Anti-Spoofing-using-DeePixBiS
  8. M8: Another CNN model trained using darknet. We have trained this model with a private dataset. This model tries to detect mobile or printed photo in the input image.

Requirements:

python ==> 3.6

pip3 install pip --upgrade

pip3 install -r requirements.txt

Usage:

python3 test.py

TODO:

  1. ?