Research map / Point clouds / Segmentation and detection
GeoFree-CoSeg: Unsupervised Point Cloud-Image Cross-Modal Co-Segmentation Without Geometric Alignment
Xin Duan, Xiabi Liu, Liyuan Pan
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 MGCM, GCSF and SFE to the basic two-branch baseline (ablation) | two-branch baseline · S3DIS (3D branch) | mIoU: 0.46 -> 0.54 | Improved |
| add MGCM, GCSF and SFE to the basic two-branch baseline (ablation) | two-branch baseline · S3DIS-Coseg (2D branch) | Jaccard: 0.39 -> 0.59 | Improved |
| add object-consistency loss alongside semantic-consistency loss (ablation) | GeoFree-CoSeg · S3DIS (3D branch) | mIoU: 0.49 -> 0.54 | Improved |
| add semantic-consistency loss alongside object-consistency loss (ablation) | GeoFree-CoSeg · S3DIS (3D branch) | mIoU: 0.51 -> 0.54 | Improved |
| add channel-level and global-level correlation to patch-level correlation in MGCM (ablation) | 2D branch · S3DIS-Coseg | Jaccard: 0.38 -> 0.52 | Improved |
| add 3D-branch cross-modal semantics to the 2D branch (ablation) | 2D branch · S3DIS-Coseg | Jaccard: 0.52 -> 0.59 | Improved |
| use Top-K graph sparsification with K around 20 to 35 instead of very small or large K (ablation) | graph-based common semantic filtering · S3DIS / ScanObjectNN | mIoU | Mixed |
| add cross-modal semantics to the 3D co-segmentation branch (GeoFree-CoSeg vs LogoSP) | 3D branch · S3DIS | mIoU: 0.48 -> 0.54 | Improved |
| add 3D cross-modal guidance to 2D co-segmentation (vs SCoSPARC) | 2D branch · ScanObjectNN-Coseg | Jaccard: 0.48 -> 0.53 | Improved |
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.
- LogoSP — LogoSP: Local-global Grouping of Superpoints for Unsupervised Semantic Segmentation of 3D Point Clouds (CVPR 2025) · better on 2 of 2 shared results
- GrowSP — GrowSP: Unsupervised Semantic Segmentation of 3D Point Clouds (CVPR 2023) · 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
- UniSeg: A Unified Multi-Modal LiDAR Segmentation Network and the OpenPCSeg Codebase (ICCV 2023 · similarity 0.89)
- Rethinking Few-shot 3D Point Cloud Semantic Segmentation (CVPR 2024 · similarity 0.87)
- FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimators (ICLR 2024 · similarity 0.86)
- LogoSP: Local-global Grouping of Superpoints for Unsupervised Semantic Segmentation of 3D Point Clouds (CVPR 2025 · similarity 0.85)
- How Do Images Align and Complement LiDAR? Towards a Harmonized Multi-modal 3D Panoptic Segmentation (ICML 2025 · similarity 0.85)
- FreePoint: Unsupervised Point Cloud Instance Segmentation (CVPR 2024 · similarity 0.85)
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.
