Research map / Point clouds / Registration and matching
UMERegRobust – Universal Manifold Embedding Compatible Features for Robust Point Cloud Registration
Yuval Haitman, Amit Efraim, Joseph M Francos
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 |
|---|---|---|---|
| replace Horn/RANSAC-style point-coordinate estimation with UME-based estimation and feature-correlation hypothesis selection | UMERegRobust · KITTI (strict precision) | registration recall (%): 78.6 -> 87.8 | Improved |
| use UMERegRobust on large-rotation scans versus GCL | UMERegRobust · RotKITTI (strict precision) | registration recall (%): 28.8 -> 73.3 | Improved |
| use UMERegRobust versus GeoTransformer on large rotations | UMERegRobust · RotKITTI (strict precision) | registration recall (%): 50.1 -> 73.3 | Improved |
| use UMERegRobust versus GCL on low-overlap large-translation scans | UMERegRobust · LoKITTI (normal precision) | registration recall (%): 72.3 -> 59.3 | Hurt |
| use UMERegRobust versus GCL on low-density LiDAR low overlap | UMERegRobust · LonuScenes (strict precision) | registration recall (%): 5.6 -> 56.3 | Improved |
| use UMERegRobust on indoor 3DMatch versus recent baselines | UMERegRobust · 3DMatch | registration recall | No clear effect |
| add UME-compatible losses to feature extractor training (ablation) | UMERegRobust coloring module · KITTI and RotKITTI | registration recall | Improved |
| add Sampling Equalizer Module to non-UME coloring (ablation) | UMERegRobust coloring module · KITTI and RotKITTI | registration recall | Improved |
| add Sampling Equalizer Module to GCL baseline (ablation) | GCL · KITTI and RotKITTI (normal precision) | registration recall | Mixed |
| match neighborhoods with UME descriptor and d_pF distance instead of point-wise l2 on UME point descriptors (ablation) | local UME descriptors · keypoint matching PR curve | matching precision: 30% deterioration with l2 | Improved |
| apply ICP refinement after registration (ablation) | UMERegRobust · outdoor benchmarks | registration recall: 2% to 4% degradation without ICP | 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 7 of 7 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
- Distinctiveness Oriented Positional Equilibrium for Point Cloud Registration (ICCV 2021 · similarity 0.85)
- RGGT: A Generative-Prior-Guided Transformer for Unified Rigid and Non-Rigid Point Cloud Registration (ICML 2026 · similarity 0.85)
- Dual Manifold Regularization Steered Robust Representation Learning for Point Cloud Analysis (AAAI 2025 · similarity 0.84)
- Mining and Transferring Feature-Geometry Coherence for Unsupervised Point Cloud Registration (NeurIPS 2024 · similarity 0.84)
- PSReg: Prior-guided Sparse Mixture of Experts for Point Cloud Registration (AAAI 2025 · similarity 0.84)
- SPEAL: Skeletal Prior Embedded Attention Learning for Cross-Source Point Cloud Registration (AAAI 2024 · similarity 0.84)
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.
