Reproduce “TactoHand: Learning Geometry-Based Dense Hand-Surface Interaction Fields from Video” 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
Ziheng Xi, Weibo Li, Jiayi Yu, Sipeng Zheng, Wanpeng Zhang, Haiweng Xu, Hao Luo, Chuan Mao and 1 more
TactoHand predicts contact, proximity, and force maps over a hand surface from monocular video. It learns from geometric labels and object motion, treating motion-derived forces as regularized estimates, not measurements. On held-out HOCap, temporal contact Macro-F1 is 50.5 versus HACO’s 18.0, using geometry-derived references rather than tactile ground truth in evaluation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.