Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Robust Outlier Rejection for 3D Registration With Variational Bayes
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 spatial-consistency non-local features with variational Bayesian non-local features (ablation) | PointDSC · 3DMatch (FPFH) | RR: 77.76 -> 80.53 | Helped |
| replace spatial-consistency non-local features with variational Bayesian non-local features (ablation) | PointDSC · 3DLoMatch (FCGF, 500 points) | RR: 36.8 -> 42.2 | Helped |
| add voting-based inlier search to VBNonlocal features (ablation) | PointDSC with VBNonlocal feat · 3DLoMatch (FCGF, 500 points) | RR: 42.2 -> 44.7 | Helped |
| add conservative seed selection (ablation) | VBNonlocal feat + Vote · 3DLoMatch (FCGF, 500 points) | RR: 44.7 -> 47.2 (+2.5%) | Helped |
| add conservative seed selection (ablation) | VBNonlocal feat + Vote · 3DLoMatch (FCGF, 250 points) | RR: 31.1 -> 34.5 (+3.4%) | Helped |
| add classification loss to intermediate SCNonlocal features (ablation) | PointDSC · 3DMatch and 3DLoMatch | RR | Mixed |
| vary non-local iteration count from 12 to 6 or 9 (ablation) | VBReg · 3DMatch and 3DLoMatch | RR | Mixed |
| vary random feature dimension (ablation) | VBReg · 3DMatch and 3DLoMatch | RR | Mixed |
| use VBReg rather than PointDSC | 3D registration · 3DMatch (FCGF) | RR: 92.42 -> 93.53 | Helped |
| use VBReg rather than PointDSC | 3D registration · 3DMatch (FPFH) | RR: 77.51 -> 82.75 | Helped |
| use VBReg rather than PointDSC | 3D registration · 3DLoMatch (FCGF, 250 points) | RR: 26.7 -> 34.5 | Hurt |
Closest papers by meaning
- A robust inlier identification algorithm for point cloud registration via $\mathbf{\ell_0}$-minimization (NeurIPS 2024 · similarity 0.89)
- Progressive Correspondence Regenerator for Robust 3D Registration (CVPR 2025 · similarity 0.88)
- Scalable 3D Registration via Truncated Entry-wise Absolute Residuals (CVPR 2024 · similarity 0.87)
- Inlier Confidence Calibration for Point Cloud Registration (CVPR 2024 · similarity 0.86)
- Locality Preserving Refinement for Shape Matching with Functional Maps (AAAI 2024 · similarity 0.84)
- Registration is a Powerful Rotation-Invariance Learner for 3D Anomaly Detection (NeurIPS 2025 · 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.