Reproduce “Clutt3R-Seg: Sparse-view 3D Instance Segmentation for Language-grounded Grasping in Cluttered Scenes” 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
Jeongho Noh, Tai Hyoung Rhee, Eunho Lee, Jeongyun Kim, Sunwoo Lee, Ayoung Kim
Clutt3R-Seg segments objects in clutter from a few camera views. It organizes noisy masks into containment hierarchies, groups evidence across views using spatial and semantic cues, and conditionally substitutes parent masks for residual fragments. On GraspClutter6D’s difficult subset with eight views, it reaches AP@25 of 61.66 versus 27.95 for MaskClustering; this measures segmentation, not robot success or recovery of hidden objects.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.