Reproduce “VoteHMR: Occlusion-Aware Voting Network for Robust 3D Human Mesh Recovery from Partial Point Clouds” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Guanze Liu, Yu Rong, Lu Sheng
VoteHMR estimates a parametric mesh from a single partial point cloud by predicting body-part labels and joint-directed votes, clustering joint evidence, and completing occluded joints before regressing SMPL parameters. On SURREAL, it reports 20.2 mm PVE versus 21.2 mm for Wang et al.’s single-frame variant; on Berkeley MHAD, its 51.76 mm Chamfer Distance measures fit to partial points, not unseen-body accuracy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.