Reproduce “CoLa3D: Composable Latent 3D Decomposition” 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
Xudong Cai, Yongcai Wang, Deying Li, Chuanxia Zheng
The authors introduce CoLa3D, which extracts editable object meshes from a shared scene representation, preserving their positions. It can use image or mesh inputs and mask, point, or learned queries. On filtered 3D-FRONT scenes, CoLa3D (Free) achieves scene Chamfer Distance of 0.0075, versus 0.0104 for SAM 3D. These results apply to selected scenes with 2–8 visible objects.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.