Research map / 3D vision / 3D generation
SGS-3D: High-Fidelity 3D Instance Segmentation via Reliable Semantic Mask Splitting and Growing
Chaolei Wang, Yang Luo, Jing Du, Siyu Chen, Yiping Chen, Ting Han
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add co-occurrence mask filtering (ablation) | SGS-3D · ScanNet200 (GT depth) | mAP: 25.9 -> 27.8 | Improved |
| add spatial continuity splitting on top of CMF (ablation) | SGS-3D · ScanNet200 (GT depth) | mAP: 27.8 -> 30.1 | Improved |
| add feature-guided growing on top of CMF+SCP (ablation) | SGS-3D · ScanNet200 (GT depth) | mAP: 30.1 -> 34.3 | Improved |
| lower input images from 10% to 2.5% (ablation) | SGS-3D · ScanNet200 (GT depth) | mAP: 34.3 -> 31.8 | Hurt |
| replace GD-SAM with SAM as 2D model (ablation) | SGS-3D · ScanNet200 (GT depth) | mAP: 34.3 -> 32.3 | Hurt |
| lower co-occurrence threshold to 0.1 (ablation) | SGS-3D · ScanNet200 (GT depth) | mAP: 34.3 -> 32.0 | Hurt |
| vary co-occurrence threshold 0.4 to 0.2 (ablation) | SGS-3D · ScanNet200 (GT depth) | mAP: 33.8, 34.0, 34.3 | No clear effect |
| remove 50% of foreground masks (ablation) | SGS-3D · ScanNet200 (GT depth) | mAP: 34.3 -> 29.6 | Hurt |
| use predicted-visibility instead of GT depth | SGS-3D · ScanNet200 | mAP: 34.3 -> 30.5 | Hurt |
Later papers that compared with this method
1 later accepted paper reports SGS-3D as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- MV3DIS: Multi-View Mask Matching via 3D Guides for Zero-Shot 3D Instance Segmentation (CVPR 2026 · better on 3 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- SAM3D — SAM 3D: 3Dfy Anything in Images (CVPR 2026) · better on 1 of 1 shared result
- SAM2Object — SAM2Object: Consolidating View Consistency via SAM2 for Zero-Shot 3D Instance Segmentation (CVPR 2025) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Geometry-Guided Semantic Reconstruction for 3D Instance Segmentation (NeurIPS 2026 · similarity 0.90)
- Any3DIS: Class-Agnostic 3D Instance Segmentation by 2D Mask Tracking (CVPR 2025 · similarity 0.90)
- SAI3D: Segment Any Instance in 3D Scenes (CVPR 2024 · similarity 0.89)
- Segment Any 3D Object with Language (ICLR 2025 · similarity 0.89)
- Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance (CVPR 2024 · similarity 0.89)
- SAM2Object: Consolidating View Consistency via SAM2 for Zero-Shot 3D Instance Segmentation (CVPR 2025 · similarity 0.89)
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.
