Research map / Point clouds / Registration and matching
Fast and Robust Iterative Closest Point
Juyong Zhang, Yuxin Yao, Bailin Deng
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 Anderson acceleration to ICP (Fast ICP vs ICP) | point-to-point ICP · Synthetic partial overlap (Bimba) | time (s): 0.33 -> 0.12 | Improved |
| add Anderson acceleration to ICP (Fast ICP vs ICP) | point-to-point ICP · 3DMatch (80-100% overlap) | time (s): 0.06 -> 0.05 | Improved |
| add Anderson acceleration to ICP (Fast ICP vs ICP) | point-to-point ICP · RGB-D SLAM (xyz) | time (s): 0.23 -> 0.14 | Improved |
| add Anderson acceleration to ICP (Fast ICP vs ICP) | point-to-point ICP · Synthetic partial overlap (Bimba) | average RMSE (x1e-3): 68 -> 68 | No clear effect |
| use Welsch robust loss with Anderson acceleration instead of ICP | ICP · Synthetic partial overlap (Bimba) | average RMSE (x1e-3): 68 -> 0.87 | Improved |
| use Welsch robust loss instead of ICP | ICP · ETH laser (Wood in winter) | average RMSE (x1e-3): 8.8 -> 0.4 | Improved |
| use robust ICP instead of sparse ICP for speed | sparse ICP · Synthetic partial overlap (Bimba) | time (s): 37.90 -> 0.96 | Improved |
| use robust ICP instead of sparse ICP for speed | sparse ICP · 3DMatch (80-100% overlap) | time (s): 9.43 -> 0.17 | Improved |
| use robust ICP instead of sparse ICP | sparse ICP · 3DMatch (80-100% overlap) | average RMSE (x1e-2): 0.75 -> 0.76 | No clear effect |
| use robust point-to-plane ICP instead of point-to-point robust ICP | Robust ICP · Synthetic partial overlap (Bunny) | average RMSE (x1e-3): 0.85 -> 0.79 | Improved |
| use robust ICP instead of DGR | DGR · RGB-D SLAM (xyz) | average RMSE (x1e-2): 0.6 -> 0.5 | Mixed |
| use robust ICP-l instead of symmetric ICP | symmetric ICP · 3DMatch (80-100% overlap) | average RMSE (x1e-2): 0.36 -> 0.71 | Hurt |
Closest papers by meaning
- Go-ICP: A Globally Optimal Solution to 3D ICP Point-Set Registration (TPAMI 2016 · similarity 0.92)
- An Approach to Robust ICP Initialization (TPAMI 2023 · similarity 0.89)
- GRICP: Granular-Ball Iterative Closest Point with Multikernel Correntropy for Point Cloud Fine Registration (AAAI 2025 · similarity 0.88)
- Scalable 3D Registration via Truncated Entry-wise Absolute Residuals (CVPR 2024 · similarity 0.85)
- A robust inlier identification algorithm for point cloud registration via $\mathbf{\ell_0}$-minimization (NeurIPS 2024 · similarity 0.84)
- Convex Relaxation for Robust Vanishing Point Estimation in Manhattan World (CVPR 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.
