Reproduce “PruneGround: Plug-and-play Spatial Pruning for 3D Visual 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

Duc Cao Dinh, Khai Le-Duc, Florent Draye, Chris Ngo, Terry Jingchen Zhang, Bernhard Schölkopf, Zhijing Jin

PruneGround grounds natural-language descriptions in 3D scenes by narrowing the scene region, using rendered local views to simplify and augment the description, and predicting one target box with an adapted SpatialLM-based model. On ScanRefer, it reports 56.1% [email protected], versus 53.9% for VGMamba. It leads in nine of ten Nr3D/Sr3D settings, but not Sr3D view-dependent.

Summary by alphaXiv.

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