Learning motion priors for 4D human body capture in 3D scenes (Talk)
It is challenging to recover realistic human-scene interactions and high-quality human motions while dealing with occlusions and partial views with a monocular RGB(D) camera. We address this problem by learning motion smoothness and infilling priors from the large scale mocap dataset AMASS, to reduce the jitters, and handle contacts and occlusions, respectively. Furthermore, we combine them into a multi-stage optimization pipeline for the high quality 4D human capture in complex 3D scenes.
Biography: Siwei Zhang is currently a second year PhD student at Computer Vision and Learning Group (VLG) at ETH. Before that, she obtained her master degree in Electrical Engineering and Information Technology from ETH, and bachelor degree in Automation from Tsinghua University. Her research interest includes human-scene interactions, 3D human pose estimation, etc.
