mlsanityData to 14 Oct 2026

NoiseSDF2NoiseSDF: Learning Clean Neural Fields from Noisy Supervision

arXiv 2507.13595 ↗First posted 18 Jul 2025cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Reconstructing accurate implicit surface representations from point clouds remains a challenging task, particularly when data is captured using low-quality scanning devices. These point clouds often contain substantial noise, leading to inaccurate surface reconstructions. Inspired by the Noise2Noise paradigm for 2D images, we introduce NoiseSDF2NoiseSDF, a novel method designed to extend this concept to 3D neural fields. Our approach enables learning clean neural SDFs from noisy point clouds through noisy supervision by minimizing the MSE loss between noisy SDF representations, allowing the network to implicitly denoise and refine surface estimations. We evaluate the effectiveness of NoiseSDF2NoiseSDF on benchmarks, including the ShapeNet, ABC, Famous, and Real datasets. Experimental results demonstrate that our framework significantly improves surface reconstruction quality from noisy inputs.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 49 of its 60 references that are in the map. First posted on arXiv 18 Jul 2025.

Builds on

49 of 60 references.

References · 49

  1. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  2. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,611
  3. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,897
  4. 2019DeepSDF: Learning Continuous Signed Distance Functions for Shape Representationcited by 2,689
  5. 2018Occupancy Networks: Learning 3D Reconstruction in Function Spacecited by 2,041
  6. 2018Learning Implicit Fields for Generative Shape Modelingcited by 1,060
  7. 2018Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labelscited by 1,059
  8. 2018Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labelscited by 997
  9. 2020Convolutional Occupancy Networkscited by 761
  10. 2020Implicit Geometric Regularization for Learning Shapescited by 710
  11. 2018Noise2Noise: Learning Image Restoration without Clean Datacited by 533
  12. 20233DShape2VecSet: A 3D Shape Representation for Neural Fields and Generative Diffusion Modelscited by 458
37 more
  1. 2019SAL: Sign Agnostic Learning of Shapes from Raw Datacited by 390
  2. 2018Noise2Void - Learning Denoising from Single Noisy Imagescited by 364
  3. 2020Implicit Functions in Feature Space for 3D Shape Reconstruction and Completioncited by 343
  4. 2018ABC: A Big CAD Model Dataset For Geometric Deep Learningcited by 341
  5. 2020Deep Local Shapes: Learning Local SDF Priors for Detailed 3D Reconstructioncited by 321
  6. 2019Noise2Self: Blind Denoising by Self-Supervisioncited by 276
  7. 2021Shape As Points: A Differentiable Poisson Solvercited by 187
  8. 2021Neighbor2Neighbor: Self-Supervised Denoising from Single Noisy Imagescited by 173
  9. 2020Neural-Pull: Learning Signed Distance Functions from Point Clouds by Learning to Pull Space onto Surfacescited by 147
  10. 2019Noisier2Noise: Learning to Denoise from Unpaired Noisy Datacited by 126
  11. 2020Points2Surf: Learning Implicit Surfaces from Point Cloud Patchescited by 122
  12. 2020SALD: Sign Agnostic Learning with Derivativescited by 115
  13. 2022ShapeFormer: Transformer-based Shape Completion via Sparse Representationcited by 106
  14. 2023Neural Kernel Surface Reconstructioncited by 104
  15. 20223DILG: Irregular Latent Grids for 3D Generative Modelingcited by 93
  16. 2022POCO: Point Convolution for Surface Reconstructioncited by 92
  17. 2020PatchNets: Patch-Based Generalizable Deep Implicit 3D Shape Representationscited by 81
  18. 2019SG-NN: Sparse Generative Neural Networks for Self-Supervised Scene Completion of RGB-D Scanscited by 77
  19. 2021DiGS : Divergence guided shape implicit neural representation for unoriented point cloudscited by 77
  20. 2019Total Denoising: Unsupervised Learning of 3D Point Cloud Cleaningcited by 58
  21. 2020Noise2Same: Optimizing A Self-Supervised Bound for Image Denoisingcited by 56
  22. 2019SSRNet: Scalable 3D Surface Reconstruction Networkcited by 35
  23. 2023Neural-Singular-Hessian: Implicit Neural Representation of Unoriented Point Clouds by Enforcing Singular Hessiancited by 35
  24. 2020Sign-Agnostic Implicit Learning of Surface Self-Similarities for Shape Modeling and Reconstruction from Raw Point Cloudscited by 30
  25. 2023IterativePFN: True Iterative Point Cloud Filteringcited by 28
  26. 2023Learning Signed Distance Functions from Noisy 3D Point Clouds via Noise to Noise Mappingcited by 27
  27. 2022Latent Partition Implicit with Surface Codes for 3D Representationcited by 21
  28. 2024LaGeM: A Large Geometry Model for 3D Representation Learning and Diffusioncited by 16
  29. 2024Fast Learning of Signed Distance Functions from Noisy Point Clouds via Noise to Noise Mappingcited by 15
  30. 2021Neural-IMLS: Self-supervised Implicit Moving Least-Squares Network for Surface Reconstructioncited by 13
  31. 2022PointCaM: Cut-and-Mix for Open-Set Point Cloud Learningcited by 11
  32. 2022Semi-signed prioritized neural fitting for surface reconstruction from unoriented point cloudscited by 8
  33. 2024Implicit Filtering for Learning Neural Signed Distance Functions from 3D Point Cloudscited by 7
  34. 2024Inferring Neural Signed Distance Functions by Overfitting on Single Noisy Point Clouds through Finetuning Data-Driven based Priorscited by 7
  35. 2025Noise2Score3D: Tweedie's Approach for Unsupervised Point Cloud Denoisingcited by 4
  36. 2024PPSURF: Combining Patches and Point Convolutions for Detailed Surface Reconstructioncited by 3
  37. 2024NumGrad-Pull: Numerical Gradient Guided Tri-plane Representation for Surface Reconstruction from Point Cloudscited by 2

Built on by

Not cited yet.