Research map / Point clouds / Registration and matching
CoFiNet: Reliable Coarse-to-fine Correspondences for Robust PointCloud Registration
Hao Yu, Fu Li, Mahdi Saleh, Benjamin Busam, Slobodan Ilic
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 fine-level refinement after coarse matching (ablation) | CoFiNet · 3DMatch, 2500 samples | Registration Recall (%): 79.6 -> 88.9 | Improved |
| add fine-level refinement after coarse matching (ablation) | CoFiNet · 3DLoMatch, 2500 samples | Registration Recall (%): 41.2 -> 66.2 | Improved |
| use overlap-ratio weighting instead of binary mask at coarse level (ablation) | CoFiNet · 3DMatch, 2500 samples | Registration Recall (%): 87.4 -> 88.9 | Improved |
| use overlap-ratio weighting instead of binary mask at coarse level (ablation) | CoFiNet · 3DLoMatch, 2500 samples | Registration Recall (%): 61.5 -> 66.2 | Improved |
| add density-adaptive matching module (ablation) | CoFiNet · 3DMatch, 2500 samples | Registration Recall (%): 88.3 -> 88.9 | Improved |
| add density-adaptive matching module (ablation) | CoFiNet · 3DLoMatch, 2500 samples | Registration Recall (%): 65.1 -> 66.2 | Improved |
| raise coarse threshold tau_c from 0.05 to 0.15 (ablation) | CoFiNet · 3DMatch, 2500 samples | Inlier Ratio (%): 49.4 -> 55.5 | Improved |
| raise coarse threshold tau_c from 0.05 to 0.15 (ablation) | CoFiNet · 3DMatch, 2500 samples | Registration Recall (%): 87.4 -> 85.5 | Hurt |
| raise coarse threshold tau_c from 0.05 to 0.15 (ablation) | CoFiNet · 3DLoMatch, 2500 samples | Registration Recall (%): 62.8 -> 58.9 | Hurt |
| add minimum coarse correspondence count tau_m=200 (with tau_c=0.20) (ablation) | CoFiNet · 3DLoMatch, 2500 samples | Registration Recall (%): 58.9 -> 66.2 | Improved |
| add minimum coarse correspondence count tau_m=200 (with tau_c=0.20) (ablation) | CoFiNet · 3DMatch, 2500 samples | Inlier Ratio (%): 55.5 -> 51.2 | Hurt |
| sample fewer correspondences (5000 to 250) (ablation) | CoFiNet · 3DMatch | Inlier Ratio (%): 49.8 -> 52.2 | Improved |
Later papers that compared with this method
11 later accepted papers report CoFiNet as a baseline in their results tables; 11 of them report a better number on most of the results they share with it.
- PointCNN++: Performant Convolution on Native Points (CVPR 2026 · better on 2 of 2 shared results)
- Register Anything Model for Generalizable and Robust Point Cloud Registration (NeurIPS 2026 · better on 1 of 1 shared result)
- Cross3DReg: A Real-world Cross-source Point Cloud Registration Benchmark (IJCV 2026 · better on 1 of 1 shared result)
- RARE: Refine Any Registration of Pairwise Point Clouds via Zero-Shot Learning (ICCV 2025 · better on 1 of 1 shared result)
- PointRegGPT: Boosting 3D Point Cloud Registration using Generative Point-Cloud Pairs for Training (ECCV 2024 · better on 1 of 1 shared result)
- UMERegRobust – Universal Manifold Embedding Compatible Features for Robust Point Cloud Registration (ECCV 2024 · better on 1 of 1 shared result)
- GeoTransformer: Fast and Robust Point Cloud Registration With Geometric Transformer (TPAMI 2023 · better on 8 of 8 shared results)
- Density-invariant Features for Distant Point Cloud Registration (ICCV 2023 · better on 2 of 3 shared results)
3 more not listed.
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
- Multi-scale Consistency for Robust 3D Registration via Hierarchical Sinkhorn Tree (NeurIPS 2024 · similarity 0.89)
- Dual Focus-Attention Transformer for Robust Point Cloud Registration (CVPR 2025 · similarity 0.88)
- Learning Instance-Aware Correspondences for Robust Multi-Instance Point Cloud Registration in Cluttered Scenes (CVPR 2024 · similarity 0.88)
- PointMC: Multi-instance Point Cloud Registration based on Maximal Cliques (ICML 2024 · similarity 0.88)
- Mining and Transferring Feature-Geometry Coherence for Unsupervised Point Cloud Registration (NeurIPS 2024 · similarity 0.88)
- A Consistency-Aware Spot-Guided Transformer for Versatile and Hierarchical Point Cloud Registration (NeurIPS 2024 · similarity 0.87)
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.
