Reproduce “FactorizedHMR: A Hybrid Framework for Video Human Mesh Recovery” 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
Patrick Kwon, Chen Chen
FactorizedHMR estimates human 3D body motion from video by fixing a deterministic torso, shape, and camera-space root, then probabilistically completing non-torso articulation and world motion. On 3DPW-XOCC, it records 66.3 MPJPE versus 76.2 for GENMO; a controlled 3DPW comparison also favors the split, 56.2 versus 58.4. These results assess recovery against one reference motion, not posterior calibration or sample diversity.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.