cleardusk / 3DDFA_V2

The official PyTorch implementation of Towards Fast, Accurate and Stable 3D Dense Face Alignment, ECCV 2020.
MIT License
2.91k stars 515 forks source link

UV Question #110

Open alexrichardson21 opened 3 years ago

alexrichardson21 commented 3 years ago

I was confused by the .obj format. Most face-mesh projects like MediaPipe and ARCore output .obj like:

v 0.000252 -0.042638 0.063704 v 0.000902 -0.022062 0.079183 v 0.000948 -0.028589 0.063385 . . . vt 0.427942 0.304722 vt 0.526878 0.295374 vt 0.444832 0.269206 . . . f 174/43 156/119 134/220 f 247/335 34/252 8/399 f 383/124 399/59 363/216

I was able to get the vt values thru utils/uv.py but I'm confused on how to get the 6 face values instead of just 3 face values. i.e.

f 2 131 1 to f 2/? 131/? 1/?

cleardusk commented 2 years ago

I think you may need to parse BFM_UV.mat.