Face Anti-Spoofing简记2-GFA-CNN

Face Anti-Spoofing简记2-Learning Generalizable and Identity-Discriminative Representations for Face Anti-Spoofing

创新点

  1. Total Pairwise Confusion (TPC) loss
  2. GFA-CNN

总结

网络结构如图所示
Face Anti-Spoofing简记2-GFA-CNN_第1张图片
人脸识别和人脸防伪的多任务网络。
总的loss:
Face Anti-Spoofing简记2-GFA-CNN_第2张图片

TPC loss:
灵感来自PC loss。
Face Anti-Spoofing简记2-GFA-CNN_第3张图片

很明了

Fast Domain Adaptation:
Domain shift 问题

最后

TPC loss不是很理解。作者说:balance the contribution of each spoof pattern, preventing the PA representations from overfitting to dataset-specific spoof patterns.

你可能感兴趣的:(Face,Anti-Spoofing简记,Face,Anti-Spoofing简记)