Reproduce “HaNDF: Object-Conditioned Geometric Neural Hand Distance Fields” 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
Zuhao Liu, Melih Darcan, Zhengdi Yu, Riza Alp Guler, Tolga Birdal
HaNDF learns an object-conditioned distance field over hand rotations to refine poses while keeping joint rotations on their valid manifold. It uses object shape to identify plausible poses. On HOI4D denoising at noise level 0.8 rad, it reduces rotation error to 5.52° versus 7.16° for an unconditional baseline, but relies on reasonably accurate object geometry.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.