Reproduce “BridgeAfford: Bridging Query-Target Gap for Generalizable 3D Affordance Grounding” 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
Yijie Su, Yakun Zhu, Yi Bin, Pengpeng Zeng, Liang Chen, Guoqing Wang, Jingkuan Song, Heng Tao Shen
BridgeAfford combines geometric alignment of similar interaction regions across objects with visual guidance from rendered views to localize regions for language instructions in point clouds. On LASO’s unseen object–affordance pairs, it scores 21.8 aIoU (average intersection over union, measuring overlap between predicted and annotated regions), versus 18.7 for GLANCE. This supports improved localization on held-out pairs, while wrap-grasping remains difficult.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.