Research map / Point clouds / Registration and matching
Correspondence-Free Fast and Robust Spherical Point Pattern Registration
Anik Sarker, Alan T. Asbeck
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 |
|---|---|---|---|
| initialize FRS with SPMC (hybrid SPMC+FRS instead of FRS alone) (ablation) | FRS · Robust Vector Alignment Dataset A2 | rotation error | Improved |
| increase outlier fraction from B1 to B7 (ablation) | SPMC · Robust Vector Alignment Dataset A5 | median angular error (deg): 0.24 -> 11.32 | Hurt |
| use CASE spherical embedding instead of EGI (Comp2Comp) (ablation) | SPMC · ModelNet40 Comp2Comp (no corr., no noise) | rotation error (deg): 1.3 -> 0.029 | Improved |
| add 0.01 Gaussian noise to point clouds (ablation) | TEASER++ with FPFH · ModelNet40 Comp2Comp | median rotation error | Hurt |
| provide 10% true correspondences (ablation) | TEASER++ with FPFH · ModelNet40 Comp2Comp | median rotation error | No clear effect |
| replace FPFH+QUASAR with SPMC+FRS | spherical pattern registration · Robust Vector Alignment Dataset | runtime: 23.18 s -> under 1 second | Improved |
| use SPMC+FRS instead of correspondence-based FPFH+QUASAR | spherical pattern registration · Robust Vector Alignment Dataset A3-A5 one-to-one (B1) | median angular error: FPFH+QUASAR errors exceed 100 deg vs SPMC+FRS under 1 deg overall | Improved |
| train deep models on 1024 points, test on 2500 points | PCRNet, PointNetLK, RPM-Net · ModelNet40 Comp2Comp | rotation error | Hurt |
| reduce overlap between point clouds to about 20% (ablation) | SPMC+FRS point cloud registration · 3DMatch | registration success | Hurt |
| increase clutter in spherical image up to 19.3% (ablation) | SPMC+FRS · Spherical images Img1-Img5 | median angular error (deg): 0.89 overall, 1.32 at Img5 | No clear effect |
Closest papers by meaning
- SHReg: Strictly Rotation-Equivariant Point Cloud Registration via Spherical Harmonics (ECCV 2026 · similarity 0.88)
- Fast Rotation Search with Stereographic Projections for 3D Registration (TPAMI 2016 · similarity 0.87)
- Efficient Global Point Cloud Alignment Using Bayesian Nonparametric Mixtures (CVPR 2017 · similarity 0.86)
- The Alignment of the Spheres: Globally-Optimal Spherical Mixture Alignment for Camera Pose Estimation (CVPR 2019 · similarity 0.86)
- Deterministic Point Cloud Registration via Novel Transformation Decomposition (CVPR 2022 · similarity 0.86)
- A robust inlier identification algorithm for point cloud registration via $\mathbf{\ell_0}$-minimization (NeurIPS 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.
