Research map / Point clouds / Registration and matching
Transformation Decoupling Strategy Based on Screw Theory for Deterministic Point Cloud Registration With Gravity Prior
Xinyi Li, Zijian Ma, Yinlong Liu, Walter M. X. Zimmer, Hu Cao, Feihu Zhang, Alois Knoll
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 |
|---|---|---|---|
| use the screw-theory decoupled three-stage search instead of FMP+BnB | registration with gravity prior · Synthetic, 95% outliers, 100k correspondences | median running time (s): 647.7 -> 0.097 | Improved |
| use the decoupled method instead of GROR at 1000k correspondences | registration with gravity prior · Synthetic, 95% outliers, 1000k correspondences | median running time (s): >1800 -> 1.905 | Improved |
| increase correspondences from 10k to 1000k | Ours · Synthetic, 95% outliers | median running time (s): 0.009 -> 1.905 | Mixed |
| use the decoupled method instead of FGR | registration with gravity prior · WHU-TLS | average translation error (m): 37.96 -> 0.165 | Improved |
| use the decoupled method instead of FMP+BnB | registration with gravity prior · WHU-TLS | average translation error (m): 12.51 -> 0.165 | Improved |
| raise outlier rate from 40% to 80% | Ours · Synthetic, N=2000 | running time | Improved |
| use the decoupled method instead of RANSAC at 98% outlier rate | registration with gravity prior · Synthetic, N=2000, 98% outliers | speed: approximately 3x10^4 times faster | Improved |
| use the decoupled method instead of GROR on A9 rotation error | registration with gravity prior · A9 | average rotation error (deg): 1.652 -> 0.394 | Improved |
| add FMP preprocessing to BnB (ablation) | BnB · ETH-Arch | average running time (s): 18.12 -> 7.722 | Improved |
Closest papers by meaning
- Deterministic Point Cloud Registration via Novel Transformation Decomposition (CVPR 2022 · similarity 0.93)
- A robust inlier identification algorithm for point cloud registration via $\mathbf{\ell_0}$-minimization (NeurIPS 2024 · similarity 0.90)
- Efficient and Robust Point Cloud Registration via Heuristics-Guided Parameter Search (TPAMI 2024 · similarity 0.89)
- Hunter: Exploring High-Order Consistency for Point Cloud Registration With Severe Outliers (TPAMI 2023 · similarity 0.89)
- Guaranteed Outlier Removal for Point Cloud Registration with Correspondences (TPAMI 2017 · similarity 0.88)
- Inlier Confidence Calibration for Point Cloud Registration (CVPR 2024 · similarity 0.87)
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.
