Research map / Point clouds / Shape and surface completion
Leaps and Bounds: An Improved Point Cloud Winding Number Formulation for Fast Normal Estimation and Surface Reconstruction
Chamin Hewa Koneputugodage, Dylan Campbell, Stephen Gould
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 |
|---|---|---|---|
| replace unbounded wU(q) with bounded wB(q) in sGCNO (ablation) | sGCNO · GCNO dataset (18 shapes, 3k points) | PGP / angular RMSE: 49.191 -> 99.913 PGP; 98.836 -> 12.772 RMSE | Improved |
| add bounded wB(q) to WNNC step with steepest descent (Ω1) (ablation) | WNNC · GCNO dataset (18 shapes, 3k points) | angular RMSE: 14.516 -> 10.798 | Improved |
| add bounded wB(q) to WNNC step with steepest descent (Ω0) (ablation) | WNNC · GCNO dataset (18 shapes, 3k points) | angular RMSE: 43.259 -> 21.844 | Improved |
| use Levenberg-Marquardt with bounded wB(q) in both steps (Ω1) (ablation) | WNNC · GCNO dataset (18 shapes, 3k points) | angular RMSE: 14.516 -> 10.158 | Improved |
| add bounded wB in WNNC step only, with LM and no smoothing (ablation) | WNNC LM · GCNO dataset ablation (Ω=0) | angular RMSE: 28.321 -> 11.633 | Improved |
| use wB in both gradient and WNNC steps instead of only the WNNC step (ablation) | WNNC LM · GCNO dataset ablation (Ω=0) | angular RMSE: 11.633 -> 12.515 | No clear effect |
| add smoothing width Ω to steepest descent (ablation) | WNNC SD · GCNO dataset (18 shapes, 3k points) | angular RMSE: 43.259 -> 14.516 | Improved |
| add smoothing width to LM (ablation) | WNNC LM · GCNO dataset ablation (Ω=0) | angular RMSE: LM A (Ω=0) 28.321 vs LM A (Ω=Ω0) 49.658 | Hurt |
| use bounded sGCNO (SGD) on large 100k-point shapes | sGCNO · Top20SION | mean IoU: 0.9460 vs 0.9709 (iPSR) | Hurt |
| SD-based WNNC with bounded wB in WNNC step vs original WNNC | WNNC SD · Top20SION | mean angular RMSE: 22.518 -> 21.100 | Improved |
| replace LM with SGD in sGCNO | sGCNO · GCNO dataset (18 shapes, 3k points) | time (s): 32.0 vs 9123 (GCNO) | Improved |
Closest papers by meaning
- Surface Reconstruction From Point Clouds: A Survey and a Benchmark (TPAMI 2024 · similarity 0.90)
- A Survey and Benchmark of Automatic Surface Reconstruction From Point Clouds (TPAMI 2024 · similarity 0.89)
- Consistent Normal Orientation for 3D Point Clouds via Least Squares on Delaunay Graph (CVPR 2025 · similarity 0.88)
- PLATYPUS: Progressive Local Surface Estimator for Arbitrary-Scale Point Cloud Upsampling (AAAI 2025 · similarity 0.87)
- GridFormer: Point-Grid Transformer for Surface Reconstruction (AAAI 2024 · similarity 0.87)
- Arbitrary-Scale Point Cloud Upsampling by Voxel-Based Network with Latent Geometric-Consistent Learning (AAAI 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.
