AI Paper Atlas

Research map / 3D vision / 3D generation

CoSMo3D: Open-World Promptable 3D Semantic Segmentation through LLM-Guided Canonical Spatial Modeling

Li Jin, Weikai Chen, Yujie Wang, Yingda Yin, Zeyu Hu, Runze Zhang, Keyang Luo, Shengju Qian, Xin Wang, Xueying Qin

CVPR 2026 · 3D vision › 3D generation · approach: Semantic scene completion

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add hard-negative sampling (ablation)CoSMo3D baseline (Find3D retrained) · 3DCompat canonical {Part} of {Obj.}mIoU: 36.89 -> 38.30Improved
add canonical map anchoring loss (ablation)CoSMo3D · 3DCompat canonical {Part} of {Obj.}mIoU: 38.30 -> 41.71Improved
add cross-category canonicalization of training data (ablation)CoSMo3D · 3DCompat canonical {Part} of {Obj.}mIoU: 41.71 -> 43.34Improved
add canonical box calibration loss (ablation)CoSMo3D · 3DCompat canonical {Part} of {Obj.}mIoU: 43.34 -> 47.51Improved
add cross-category canonicalization (ablation)CoSMo3D · 3DCompat rotated {Part} of {Obj.}mIoU: 42.48 -> 42.63No clear effect
retrain Find3D on the curated canonical datasetFind3D · 3DCompat-Coarse canonical {Part}mIoU: 25.36 -> 45.00Improved
use CoSMo3D instead of retrained Find3DFind3D · 3DCompat-CoarsemIoU: 36.89 -> 47.51Improved
use CoSMo3D instead of Find3D on single-word promptsFind3D · PartNet-E canonicalmIoU: PartSLIP++ 32.71 vs ours 17.19Mixed
rotate objects at test timeCoSMo3D · 3DCompat-Coarse {Part} of {Obj.}mIoU: 47.51 -> 47.74No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy