yfeng95 / PRNet

Joint 3D Face Reconstruction and Dense Alignment with Position Map Regression Network (ECCV 2018)
http://openaccess.thecvf.com/content_ECCV_2018/papers/Yao_Feng_Joint_3D_Face_ECCV_2018_paper.pdf
MIT License
4.95k stars 947 forks source link

Depth map for spoofed face #135

Open bcba25 opened 5 years ago

bcba25 commented 5 years ago

Is there a way to get a depth map for printed face ? By right it should be flat, how can I test this ? When I tried to apply PRNet it also gives a 3D shape of the face for a spoofed photo

harrismohammed commented 5 years ago

Even I have the same issue. Moreover I came across a few works on Github where they manually assume depth =0 for Print and Replay attack. I would want to know how to effectively arrive at notable depth difference between Real Image and a Print/Replay image.

punitha-valli commented 4 years ago

@bcba25 @harrismohammed @ksachdeva @cclauss @scott-vsi

I have generated the depth map, can anyone please tell me hoe to increases the GPU so I can all to use all the GPU in order to increases the speed