Reproduce “ViewStruct: View-Grounded Articulation and Connectivity Inference” 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
Wenjing Bian, Ranran Huang, Andreas Geiger
ViewStruct predicts rigid bodies, attachments, and joint constraints from RGB-D images with aligned part segmentation and depth, without requiring observed motion. It grounds predictions in a gravity-aligned, view-relative frame. On 387 PhysX-Mobility test objects, it achieves 78.2% success over all reference joints, versus 66.6% for PhysX-Omni, but scores 81.5% versus 85.3% among matched joints; ViewStruct also uses richer inputs.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.