1 paper
Zehua Wang, Changwang Mei, Huaijiang Sun +2
Existing 2D-3D lifting human pose estimation methods have achieved strong performance. But the utilization of historical pose representations across network depth was overlooked. I…