Abstract: This work addresses the problem of 3D human pose and shape estimation from monocular video. While significant progress has been made in this domain, most existing approaches focus solely on ...