Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Accelerated Gravitational Point Set Alignment With Altered Physical Laws
Vladislav Golyanik, Christian Theobalt, Didier Stricker
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 prior matches (up to 3) to BH-RGA instead of none (ablation) | BH-RGA · N500-U50 | success rate: 87% with 1 prior, 100% with 2 priors, 100% with 3 priors | Improved |
| add anchor points (high-mass points) to BH-RGA (ablation) | BH-RGA · N500-U50 | success rate: 25% -> 40% (1 anchor), 38% (2), 33% (3) | Mixed |
| increase distance threshold gamma of the Barnes-Hut tree (ablation) | BH-RGA · SINTEL sleeping2 registration | RMSE and runtime: RMSE 0.068 in 11 seconds (gamma = 0.25) vs 0.0065 in 264 seconds (gamma = 3.0) | Mixed |
| increase rotation from 5 to 24 degrees (ablation) | BH-RGA · SINTEL sleeping2 registration | runtime: overall runtime increases by the factor of two; per-iteration runtime ~40% | Hurt |
| add uniform noise (N500-U100 vs clean-500) | BH-RGA vs CPD (7 DoF) · clean-500 / N500-U100 | RMSE: CPD 0.04 vs BH-RGA 0.009 (clean) ; CPD 0.088 vs BH-RGA 0.059 (N500-U100) | Mixed |
| use BH-RGA instead of GA baseline | gravitational alignment · U256 | RMSE: 0.149 -> 0.015 | Improved |
| use BH-RGA instead of ICP under perturbation | point perturbation · U256 | RMSE: 9E-3 vs 0.015 | Hurt |
Closest papers by meaning
- Gravitational Approach for Point Set Registration (CVPR 2016 · similarity 0.90)
- Efficient Registration of High-Resolution Feature Enhanced Point Clouds (TPAMI 2018 · similarity 0.84)
- Transformation Decoupling Strategy Based on Screw Theory for Deterministic Point Cloud Registration With Gravity Prior (TPAMI 2024 · similarity 0.82)
- Certifiably Optimal Robust Angular Synchronization (NeurIPS 2026 · similarity 0.82)
- Group-Wise Point-Set Registration Based on Renyi's Second Order Entropy (CVPR 2017 · similarity 0.81)
- Provably Approximated Point Cloud Registration (ICCV 2021 · similarity 0.81)
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.
