Research map / Point clouds / Point cloud classification
Nesti-Net: Normal Estimation for Unstructured 3D Point Clouds Using Convolutional Neural Networks
Yizhak Ben-Shabat, Michael Lindenbaum, Anath Fischer
Check 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 |
|---|---|---|---|
| learn a local scale choice with the mixture-of-experts scale manager rather than use a fixed single scale (ablation) | Nesti-Net · PCPNet test set, average | RMS angle error: 23.91 -> 12.41 | Improved |
| select an expert scale rather than concatenate multiple scales (ablation) | Nesti-Net · PCPNet test set, average | RMS angle error: 16.11 -> 12.41 | Improved |
| learn scale selection without noise-level labels rather than switch using a provided noise level (ablation) | Nesti-Net · PCPNet test set, average | RMS angle error: 13.55 -> 12.41 | Improved |
| use three scales instead of two (ablation) | Nesti-Net · PCPNet test set, average | RMS angle error: 12.97 -> 12.41 | Improved |
| use learned scale selection rather than noise-based switching (ablation) | Nesti-Net · PCPNet test set, noise σ = 0.012 | RMS angle error: 22.59 -> 22.28 | Improved |
| use learned scale selection rather than noise-based switching (ablation) | Nesti-Net · PCPNet test set, Gradient density | RMS angle error: 11.89 -> 9.00 | Improved |
| use Nesti-Net instead of the strongest reported baseline | normal estimation · PCPNet test set, Gradient density | RMS angle error: 8.49 -> 9.00 | Hurt |
| use Nesti-Net instead of PCPNet ss | normal estimation · PCPNet test set, average | PGP10: 0.7393 -> 0.8120 | Improved |
| use Nesti-Net instead of PCPNet ss | normal estimation · PCPNet test set, average | PGP5: 0.5526 -> 0.6617 | Improved |
Later papers that compared with this method
6 later accepted papers report Nesti-Net as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.
- Refine-Net: Normal Refinement Neural Network for Noisy Point Clouds (TPAMI 2022 · better on 8 of 9 shared results)
- GraphFit: Learning Multi-Scale Graph-Convolutional Representation for Point Cloud Normal Estimation (ECCV 2022 · better on 1 of 1 shared result)
- AdaFit: Rethinking Learning-Based Normal Estimation on Point Clouds (ICCV 2021 · better on 1 of 1 shared result)
- Deep Iterative Surface Normal Estimation (CVPR 2020 · better on 8 of 8 shared results)
- Geometry and Learning Co-Supported Normal Estimation for Unstructured Point Cloud (CVPR 2020 · better on 2 of 2 shared results)
- DeepFit: 3D Surface Fitting via Neural Network Weighted Least Squares (ECCV 2020 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- LinNet: Linear Network for Efficient Point Cloud Representation Learning (NeurIPS 2024 · similarity 0.88)
- An Explicit Frame Construction for Normalizing 3D Point Clouds (ICML 2024 · similarity 0.87)
- CMG-Net: Robust Normal Estimation for Point Clouds via Chamfer Normal Distance and Multi-Scale Geometry (AAAI 2024 · similarity 0.87)
- Learning Continuous Implicit Field with Local Distance Indicator for Arbitrary-Scale Point Cloud Upsampling (AAAI 2024 · similarity 0.87)
- DeepLA-Net: Very Deep Local Aggregation Networks for Point Cloud Analysis (CVPR 2025 · similarity 0.87)
- PLATYPUS: Progressive Local Surface Estimator for Arbitrary-Scale Point Cloud Upsampling (AAAI 2025 · 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.
