Reproduce “PhysSkin++: Real-Time and Generalizable Physics-Based Animation via Self-Supervised Neural Skinning” 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

Yuanhang Lei, Baofei Liu, Xingxuan Li, Tao Cheng, Zixuan Wang, Boming Zhao, Yujun Shen, Siyuan Huang and 4 more

PhysSkin++ learns continuous skinning fields from static shapes, then simulates motion in compact handle coordinates rather than full-space geometry. A shape-conditioned predictor is trained across objects without simulation trajectories or skinning annotations, using physical and structural objectives. On ShapeNet, its orthogonality metric is 0.0098 × 10⁻² versus 0.2621 × 10⁻² for Simplicits; this measures basis structure, not animation quality.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.