1 paper · 1 filter
Zhanke Wang, Chengfeng Zhao, Qing Shuai +8
We present FlowHMR, a framework for recovering physically plausible global 3D human motion from monocular video. Previous learning-based methods typically regress human motion dire…