Using the output for anti spoofing detection

cleardusk / 3DDFA_V2

The official PyTorch implementation of Towards Fast, Accurate and Stable 3D Dense Face Alignment, ECCV 2020.

MIT License

2.83k stars 505 forks source link

Using the output for anti spoofing detection #99

Open woreom opened 3 years ago

woreom commented 3 years ago

I'm looking for a model to get the depth of the video for my dataset, looking at the title I thought that the result for a real and a fake picture or video is different but with my testing of demo.py and demo_video.py I can't really see any difference, What am I missing?

quangtn266 commented 2 years ago

Hi, I think that the difference between real or fake pic is that real will have a depth map, and fake pic will haven't it or just numpy rezos. I already some papers about Face Anti-Spoofing, they usually define depth map for fake pic as numpy zeros.

Sorry if it's wrong opinion.

woreom commented 2 years ago

@quangtn266 Thanks