This task aims to solve absolute (camera-centric not root-relative) 3D human pose estimation.
<span style="color:grey; opacity: 0.6">( Image credit: RootNet )</span>