Research map / Point clouds / Point cloud classification
HGMR: Hierarchical Gaussian Mixtures for Adaptive 3D Registration
B. Eckart, K. Kim, J. Kautz
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 adaptive early-stopping scale selection to GMM-Tree level 4 (ablation) | HGMR · Stanford Lounge | angular error (deg): 0.56 -> 0.37 | Improved |
| add adaptive early-stopping scale selection to GMM-Tree level 3 (ablation) | HGMR · Stanford Lounge | angular error (deg): 0.48 -> 0.46 | Improved |
| add adaptive early-stopping scale selection to GMM-Tree level 2 (ablation) | HGMR · Stanford Lounge | angular error (deg): 0.77 -> 0.76 | No clear effect |
| add adaptive scale selection to GMM-Tree level 3 (ablation) | HGMR · Velodyne LiDAR | angular error (deg): 0.18 -> 0.15 | Improved |
| add adaptive scale selection to GMM-Tree level 4 (ablation) | HGMR · Velodyne LiDAR | angular error (deg): 0.20 -> 0.15 | Improved |
| add adaptive scale selection to GMM-Tree level 2 (ablation) | HGMR · Velodyne LiDAR | angular error (deg): 0.11 -> 0.12 | Hurt |
| add adaptive scale selection to GMM-Tree level 3 (ablation) | HGMR · Velodyne LiDAR | translation error (cm): 23.9 -> 8.8 | Improved |
| add adaptive scale selection to GMM-Tree level 4 (ablation) | HGMR · Velodyne LiDAR | translation error (cm): 29.5 -> 9.2 | Improved |
| add adaptive scale selection to GMM-Tree level 2 (ablation) | HGMR · Velodyne LiDAR | translation error (cm): 12.5 -> 10.0 | Improved |
| add adaptive scale selection to GMM-Tree level 4 (ablation) | HGMR · Stanford Lounge | speed (fps): 14.2 -> 14.5 | No clear effect |
| increase GMM-Tree max depth from level 2 to level 4 (static) | HGMR · Stanford Lounge | speed (fps): 31.3 -> 14.2 | Hurt |
| use Adaptive L3 HGMR instead of TrICP point-to-plane | HGMR · Stanford Lounge | angular error (deg): 0.54 -> 0.46 | Improved |
Closest papers by meaning
- DeepGMR: Learning Latent Gaussian Mixture Models for Registration (ECCV 2020 · similarity 0.93)
- PointGMM: A Neural GMM Network for Point Clouds (CVPR 2020 · similarity 0.91)
- PointMC: Multi-instance Point Cloud Registration based on Maximal Cliques (ICML 2024 · similarity 0.87)
- 3D Focusing-and-Matching Network for Multi-Instance Point Cloud Registration (NeurIPS 2024 · similarity 0.87)
- Mining and Transferring Feature-Geometry Coherence for Unsupervised Point Cloud Registration (NeurIPS 2024 · similarity 0.86)
- GRICP: Granular-Ball Iterative Closest Point with Multikernel Correntropy for Point Cloud Fine Registration (AAAI 2025 · similarity 0.86)
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.
