1 paper
Jianan Li, Xiao Chen, Tao Huang +1
Video data is more cost-effective than motion capture data for learning 3D character motion controllers, yet synthesizing realistic and diverse behaviors directly from videos remai…