search

JiangWenPL / multiperson

Code repository for the paper: "Coherent Reconstruction of Multiple Humans from a Single Image" in CVPR'20
https://jiangwenpl.github.io/multiperson/
376 stars 50 forks source link

Data for Adversarial Training #18

Closed nicolasugrinovic closed 3 years ago

nicolasugrinovic commented 3 years ago

Hi,

I was wondering, which data do you use for adversarial pre-training SMPL head? Is it the same as in Kanazawa paper? or do you use different data.

Thank you