1 paper
Zhiwei Liu, Xiangyu Zhu, Lu Yang +7
3D human pose and shape recovery from a monocular RGB image is a challenging task. Existing learning based methods highly depend on weak supervision signals, e.g. 2D and 3D joint l…