search

clks-wzz / PRNet-Depth-Generation

A implementaion of depth generation based on [PRNet](https://github.com/YadiraF/PRNet), which was used in the paper ***Exploiting Temporal and Depth Information for Multi-frame Face Anti-spoofing***
MIT License
95 stars 31 forks source link

Pose face issue #7

Open bushiwoQAQ opened 3 years ago

bushiwoQAQ commented 3 years ago

Thank you for your job! When I try to generate a pose face image, the following image appears:

DELSU3((Z1I8D1TBY(R0}53

How to solve the black area of depth map？

baopmessi commented 3 years ago

Same problem. How to solving it. Thanks.

bushiwoQAQ commented 3 years ago

Same problem. How to solving it. Thanks.

I think so, because the author does not use the Z-buffer algorithm, there is a problem in the priority between the self-occlusion part and the visualization part of the pose model. That is, the algorithm of the author is not suitable for the pose depth model.