1 paper
Guocun Wang, Kenkun Liu, Jing Lin +3
Existing 3D human motion generation and understanding methods often exhibit limited interpretability, restricting effective mutual enhancement between these inherently related task…