Research map / Point clouds / Registration and matching
ERNet: Efficient Non-Rigid Registration Network for Point Sequences
Guangzhao He, Yuxi Xiao, Zhen Xu, Xiaowei Zhou, Sida Peng
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 node-graph skinning instead of dense matching (ablation) | ERNet · D-FAUST test (ablation) | ATE_3D: 0.014 -> 0.011 | Improved |
| add node-graph skinning instead of dense matching (ablation) | ERNet · D-FAUST test (ablation) | delta_0.05: 0.885 -> 0.918 | Improved |
| add node-graph skinning instead of dense matching (ablation) | ERNet · D-FAUST test (ablation) | T_avg (s/frame): 4.645 -> 0.172 | Improved |
| add node-to-frame matching network phi (ablation) | ERNet · D-FAUST test (ablation) | ATE_3D: 0.022 -> 0.011 | Improved |
| add node-to-frame matching network phi (ablation) | ERNet · D-FAUST test (ablation) | delta_0.01: 0.289 -> 0.416 | Improved |
| add node-to-frame matching network phi (ablation) | ERNet · D-FAUST test (ablation) | T_avg (s/frame): 0.060 -> 0.172 | Hurt |
| add spatio-temporal refinement transformer Phi (ablation) | ERNet · D-FAUST test (ablation) | ATE_3D: 0.014 -> 0.011 | Improved |
| add spatio-temporal refinement transformer Phi (ablation) | ERNet · D-FAUST test (ablation) | delta_0.01: 0.366 -> 0.416 | Improved |
| add spatio-temporal refinement transformer Phi (ablation) | ERNet · D-FAUST test (ablation) | delta_0.05: 0.889 -> 0.918 | Improved |
| reduce inference update iterations from M=6 to M=4 (ablation) | ERNet · D-FAUST | ATE_3D: M=6 to M=4, minor performance drop | No clear effect |
| use B=256 graph nodes (ablation) | ERNet · D-FAUST | ATE_3D: B=256 chosen as best balance of accuracy and efficiency | Improved |
| replace chained per-sequence optimization baseline C-NDP with ERNet | C-NDP · D-FAUST test | ATE_3D: 0.041 -> 0.016 | Improved |
Closest papers by meaning
- Neural Deformation Graphs for Globally-Consistent Non-Rigid Reconstruction (CVPR 2021 · similarity 0.87)
- Deep Graph-Based Spatial Consistency for Robust Non-Rigid Point Cloud Registration (CVPR 2023 · similarity 0.87)
- EDFFDNet: Towards Accurate and Efficient Unsupervised Multi-Grid Image Registration (ICCV 2025 · similarity 0.87)
- Acceleration of Non-Rigid Point Set Registration With Downsampling and Gaussian Process Regression (TPAMI 2020 · similarity 0.87)
- Occlusion-aware Non-Rigid Point Cloud Registration via Unsupervised Neural Deformation Correntropy (ICLR 2025 · similarity 0.86)
- CARL: A Framework for Equivariant Image Registration (CVPR 2025 · similarity 0.85)
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.
