Reproduce “Function beyond Form: Functional Correspondence for Cross-Embodiment Dexterous Grasp Generation” 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
Bolin Zou, Wenlong Dong, Mu Ai, Chao Tang, Aoxiang Gu, Lipeng Chen, Hong Zhang
FunCo-Grasp transfers grasp knowledge across dexterous hands by aligning functional parts in shared local frames while retaining each hand’s geometry and kinematics. It uses diffusion and inverse kinematics to generate executable grasps, requiring one-time annotation but no target grasp data or fine-tuning. On four unseen simulated hands, it reaches 74.02% versus CEDex’s 66.50% under a displacement-based protocol, not a guarantee.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.