Research map / 3D vision / Depth and stereo
Least squares surface reconstruction on arbitrary domains
Dizhong Zhu, William A. P. Smith
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 2D SG filters with 3D KNN instead of smoothed central difference | normals-from-depth · 3D Ken Burns city-walking | median angular error: 23.60 -> 15.73 | Improved |
| use 2D SG filters with 3D KNN instead of smoothed central difference | normals-from-depth · 3D Ken Burns victorian-walking | median angular error: 26.38 -> 19.84 | Improved |
| use SG filters instead of forward difference | normals-from-depth · 3D Ken Burns city-walking | median angular error: 38.11 -> 15.73 | Improved |
| replace derivative and smoothing matrices with SG filters | Smith and Fang photometric stereo · DiLiGenT bear | normal median angular error: 12.7 -> 7.33 | Improved |
| replace derivative and smoothing matrices with SG filters | Smith and Fang photometric stereo · DiLiGenT buddha | normal median angular error: 26.3 -> 14.8 | Improved |
| replace derivative and smoothing matrices with SG filters | Smith and Fang photometric stereo · DiLiGenT pot 1 | normal median angular error: 20.0 -> 9.88 | Improved |
| replace derivative and smoothing matrices with SG filters | Smith and Fang photometric stereo · DiLiGenT cow | normal median angular error: 25.4 -> 23.7 | Improved |
| SG filters versus prior perspective integration method | height-from-normals · DiLiGenT depth (bear) | depth RMSE: 5.55 -> 5.42 | Improved |
| SG filters versus prior perspective integration method | height-from-normals · DiLiGenT depth (pot 1, pot 2, reading, harvest) | depth RMSE: 2.78 -> 3.07 | Mixed |
| add SG filters for smoothness regularisation instead of Laplacian | height-from-normals · synthetic peaks with Gaussian noise | RMSE | No clear effect |
Closest papers by meaning
- Neural Splines: Fitting 3D Surfaces With Infinitely-Wide Neural Networks (CVPR 2021 · similarity 0.86)
- VKSR: Scalable Kernel Surface Reconstruction Using Vecchia's Approximation (ECCV 2026 · similarity 0.86)
- FatesGS: Fast and Accurate Sparse-View Surface Reconstruction Using Gaussian Splatting with Depth-Feature Consistency (AAAI 2025 · similarity 0.85)
- FNIN: A Fourier Neural Operator-based Numerical Integration Network for Surface-from-gradients (AAAI 2025 · similarity 0.85)
- Sparse2DGS: Geometry-Prioritized Gaussian Splatting for Surface Reconstruction from Sparse Views (CVPR 2025 · similarity 0.84)
- Discontinuity-preserving Normal Integration with Auxiliary Edges (CVPR 2024 · 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.
