Research map / Point clouds / Registration and matching
RARE: Refine Any Registration of Pairwise Point Clouds via Zero-Shot Learning
Chengyu Zheng, Jin Huang, Honghua Chen, Mingqiang Wei
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 RARE refinement on top of GeoTransformer | GeoTransformer · 3DMatch | Registration Recall (%): 92.0 -> 94.5 | Improved |
| Add RARE refinement on top of FCGF | FCGF · 3DMatch | Registration Recall (%): 67.9 -> 76.2 | Improved |
| Add RARE refinement on top of Predator and CoFiNet | Predator, CoFiNet · 3DMatch | Registration Recall (%): 89.0 -> 92.7 (Predator); 89.3 -> 93.0 (CoFiNet) | Improved |
| Add Inlier Aggregation module (ablation) | RARE · 3DMatch | Registration Recall (%): 92.8 -> 94.5 | Improved |
| Add Feature Transfer module (ablation) | RARE · 3DMatch | Registration Recall (%): 93.1 -> 94.5 | Improved |
| Add PCA dimensionality reduction of diffusion features (ablation) | RARE · 3DMatch | Registration Recall (%): 94.1 -> 94.5 | No clear effect |
| Use diffusion layers [0,3,6] instead of single layers or adding layer 9 (ablation) | Depth diffusion features · 3DMatch (diffusion features only) | Registration Recall (%): 63.3 -> 76.0 (vs [0,3,6,9]) | Improved |
| Add extra-large layer 9 features (ablation) | Depth diffusion features · 3DMatch (diffusion features only) | Registration Recall (%): 76.0 -> 60.9 ([0,3,9] vs [0,3,6]) | Hurt |
| Add RARE refinement on top of GeoTransformer on KITTI | GeoTransformer · KITTI | Registration Recall (%): 99.8 -> 99.8 | No clear effect |
| Add RARE refinement on top of GeoTransformer on KITTI | GeoTransformer · KITTI | Relative Translation Error: 0.068 -> 0.064 | 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.
- GeoTransformer — GeoTransformer: Fast and Robust Point Cloud Registration With Geometric Transformer (TPAMI 2023) · better on 2 of 2 shared results
- CoFiNet — CoFiNet: Reliable Coarse-to-fine Correspondences for Robust PointCloud Registration (NeurIPS 2021) · 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
- FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimators (ICLR 2024 · similarity 0.93)
- Zero-shot RGB-D Point Cloud Registration with Pre-trained Large Vision Model (CVPR 2025 · similarity 0.92)
- Where Precision Meets Efficiency: Transformation Diffusion Model for Point Cloud Registration (AAAI 2025 · similarity 0.91)
- PosDiffNet: Positional Neural Diffusion for Point Cloud Registration in a Large Field of View with Perturbations (AAAI 2024 · similarity 0.89)
- DiffCorr: Conditional Diffusion Model with Reliable Pseudo-Label Guidance for Unsupervised Point Cloud Shape Correspondence (AAAI 2025 · similarity 0.89)
- Point Cloud Pre-training with Diffusion Models (CVPR 2024 · 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.
