Research map / Point clouds / Segmentation and detection
PointDC: Unsupervised Semantic Segmentation of 3D Point Clouds via Cross-Modal Distillation and Super-Voxel Clustering
Zisheng Chen, Hongbin Xu, Weitao Chen, Zhipeng Zhou, Haihong Xiao, Baigui Sun, Xuansong Xie, Wenxiong kang
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 multi-view 2D clustering over random 3D clustering (ablation) | random clustering · ScanNet-v2 val | mIoU: 3.86 -> 13.58 | Improved |
| add cross-modal distillation (CMD) (ablation) | PointDC · ScanNet-v2 val | mIoU: 13.58 -> 20.29 | Improved |
| add basic super-voxel clustering (ablation) | PointDC · ScanNet-v2 val | mIoU: 20.29 -> 23.64 | Improved |
| add non-parametric classifier (ablation) | PointDC · ScanNet-v2 val | mIoU: 23.64 -> 24.85 | Improved |
| add super-voxel pooling of pseudo labels (ablation) | PointDC · ScanNet-v2 val | mIoU: 24.85 -> 25.74 | Improved |
| add super-voxel pooling of pseudo labels (ablation) | PointDC · ScanNet-v2 val | accuracy: 61.79 -> 63.69 | Improved |
| add super-voxel clustering stage after CMD with DGCNN (ablation) | PointDC (DGCNN) · ScanNet-v2 val | mIoU: 11.50 -> 14.06 | Improved |
| add super-voxel clustering stage after CMD with DGCNN (ablation) | PointDC (DGCNN) · ScanNet-v2 val | accuracy: 39.65 -> 45.36 | Improved |
| PointDC versus best baseline STRL with DGCNN backbone | DGCNN · ScanNet-v2 val | unsupervised mIoU: +9.93 unsupervised mIoU and +24.84 unsupervised accuracy | Improved |
| PointDC versus next best baseline on ScanNet-v2 test | PointDC · ScanNet-v2 test | mIoU: +18.4 unsupervised mIoU | Improved |
Later papers that compared with this method
3 later accepted papers report PointDC as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- PointGS: Semantic-Consistent Unsupervised 3D Point Cloud Segmentation with 3D Gaussian Splatting (CVPR 2026 · better on 2 of 2 shared results)
- Tri-Perspective View Decomposition for Geometry-Aware Depth Completion (CVPR 2024 · better on 2 of 2 shared results)
- DeCoTR: Enhancing Depth Completion with 2D and 3D Attentions (CVPR 2024 · 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
- GrowSP: Unsupervised Semantic Segmentation of 3D Point Clouds (CVPR 2023 · similarity 0.92)
- LogoSP: Local-global Grouping of Superpoints for Unsupervised Semantic Segmentation of 3D Point Clouds (CVPR 2025 · similarity 0.91)
- P-SLCR: Unsupervised Point Cloud Semantic Segmentation via Prototypes Structure Learning and Consistent Reasoning (AAAI 2026 · similarity 0.91)
- PointGS: Semantic-Consistent Unsupervised 3D Point Cloud Segmentation with 3D Gaussian Splatting (CVPR 2026 · similarity 0.90)
- Geometry and Uncertainty-Aware 3D Point Cloud Class-Incremental Semantic Segmentation (CVPR 2023 · similarity 0.90)
- FreePoint: Unsupervised Point Cloud Instance Segmentation (CVPR 2024 · similarity 0.90)
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.
