python and Tensorflow implementation of a CNN for face anti-spoofing This project is based on " Deep Learning for Face Anti-Spoofing: An End-to-End Approach" published by Yasar Abbas Ur Rehamn, Lai Man Po and Mengyang Liu. This paper proposed a method for face anti-spoofing, which could judge whether the input image is a real human face or a fake photo/video frame. It proposed three CNN models based on VGG-11, Model A,B,C, which have realized accuracies of 84%,81%,88%. Based on these models, I made a few adjustments. And it ultimately realized accuracies of 93.74%,91.15%,92.12% of these three models, which has improved a lot.
zjz19960805/CNN-for-face-anti-spoofing
python and Tensorflow implementation of a CNN for face anti-spoofing
Python