mlsanityData to 14 Oct 2026

Toward Robust Neural Reconstruction from Sparse Point Sets

arXiv 2412.16361 ↗First posted 20 Dec 2024cs.CVcited by 5

Not on the map. Why?

Abstract (arXiv)

We consider the challenging problem of learning Signed Distance Functions (SDF) from sparse and noisy 3D point clouds. In contrast to recent methods that depend on smoothness priors, our method, rooted in a distributionally robust optimization (DRO) framework, incorporates a regularization term that leverages samples from the uncertainty regions of the model to improve the learned SDFs. Thanks to tractable dual formulations, we show that this framework enables a stable and efficient optimization of SDFs in the absence of ground truth supervision. Using a variety of synthetic and real data evaluations from different modalities, we show that our DRO based learning framework can improve SDF learning with respect to baselines and the state-of-the-art methods.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 63 of its 94 references that are in the map. First posted on arXiv 20 Dec 2024.

Builds on

63 of 94 references.

References · 63

  1. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  2. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,897
  3. 2019DeepSDF: Learning Continuous Signed Distance Functions for Shape Representationcited by 2,689
  4. 2013Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distancescited by 2,316
  5. 2020Implicit Neural Representations with Periodic Activation Functionscited by 2,238
  6. 2018Occupancy Networks: Learning 3D Reconstruction in Function Spacecited by 2,041
  7. 2021NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstructioncited by 1,712
  8. 2021Efficient Geometry-aware 3D Generative Adversarial Networkscited by 1,205
  9. 2021Volume Rendering of Neural Implicit Surfacescited by 996
  10. 2016A Point Set Generation Network for 3D Object Reconstruction from a Single Imagecited by 979
  11. 2020Convolutional Occupancy Networkscited by 761
  12. 2020Implicit Geometric Regularization for Learning Shapescited by 710
38 more
  1. 2018Pixel2Mesh: Generating 3D Mesh Models from Single RGB Imagescited by 651
  2. 2021Zero-Shot Text-Guided Object Generation with Dream Fieldscited by 513
  3. 2019BlendedMVS: A Large-scale Dataset for Generalized Multi-view Stereo Networkscited by 490
  4. 2018Generalizing to Unseen Domains via Adversarial Data Augmentationcited by 451
  5. 2017Neural 3D Mesh Renderercited by 437
  6. 2019SAL: Sign Agnostic Learning of Shapes from Raw Datacited by 390
  7. 2020Local Implicit Grid Representations for 3D Scenescited by 373
  8. 2017Certifying Some Distributional Robustness with Principled Adversarial Trainingcited by 337
  9. 2019Neural Point-Based Graphicscited by 327
  10. 2015Data-driven Distributionally Robust Optimization Using the Wasserstein Metric: Performance Guarantees and Tractable Reformulationscited by 291
  11. 2023Visual Geometry Grounded Deep Structure From Motioncited by 227
  12. 2021Shape As Points: A Differentiable Poisson Solvercited by 187
  13. 2016Distributionally Robust Stochastic Optimization with Wasserstein Distancecited by 185
  14. 2020Neural-Pull: Learning Signed Distance Functions from Point Clouds by Learning to Pull Space onto Surfacescited by 147
  15. 2015Distributionally Robust Logistic Regressioncited by 126
  16. 2020SemanticPOSS: A Point Cloud Dataset with Large Quantity of Dynamic Instancescited by 123
  17. 2020SALD: Sign Agnostic Learning with Derivativescited by 115
  18. 2018Deep Geometric Prior for Surface Reconstructioncited by 110
  19. 2018Auxiliary Tasks in Multi-task Learningcited by 108
  20. 2023Neural Kernel Surface Reconstructioncited by 104
  21. 2022POCO: Point Convolution for Surface Reconstructioncited by 92
  22. 2013Data-Driven Robust Optimizationcited by 85
  23. 2021DiGS : Divergence guided shape implicit neural representation for unoriented point cloudscited by 77
  24. 2021Deep Implicit Moving Least-Squares Functions for 3D Reconstructioncited by 67
  25. 2015Distributional Smoothing with Virtual Adversarial Trainingcited by 61
  26. 2022Reconstructing Surfaces for Sparse Point Clouds with On-Surface Priorscited by 60
  27. 2021Neural Fields as Learnable Kernels for 3D Reconstructioncited by 58
  28. 2022Surface Reconstruction from Point Clouds by Learning Predictive Context Priorscited by 54
  29. 2020Neural Splines: Fitting 3D Surfaces with Infinitely-Wide Neural Networkscited by 52
  30. 2020Meshing Point Clouds with Predicted Intrinsic-Extrinsic Ratio Guidancecited by 39
  31. 2021Phase Transitions, Distance Functions, and Implicit Neural Representationscited by 38
  32. 2021Sinkhorn Distributionally Robust Optimizationcited by 38
  33. 2023NeuSurf: On-Surface Priors for Neural Surface Reconstruction from Sparse Input Viewscited by 37
  34. 2022A Unified Wasserstein Distributional Robustness Framework for Adversarial Trainingcited by 36
  35. 2020Implicit Feature Networks for Texture Completion from Partial 3D Datacited by 31
  36. 2022Iterative Poisson Surface Reconstruction (iPSR) for Unoriented Pointscited by 31
  37. 2023Unsupervised Inference of Signed Distance Functions from Single Sparse Point Clouds without Learning Priorscited by 28
  38. 2020Dynamic Plane Convolutional Occupancy Networkscited by 24

50 most cited of 63 All 63 (JSON)

Built on by

Cited by year

Citing papers per year: 2025: 5.

5 papers.

  1. 2025Sparfels: Fast Reconstruction from Sparse Unposed Imagerycited by 8
  2. 2025SparSplat: Fast Multi-View Reconstruction with Generalizable 2D Gaussian Splattingcited by 7
  3. 2025Anti-Aliased 2D Gaussian Splattingcited by 7
  4. 2025TextureSplat: Per-Primitive Texture Mapping for Reflective Gaussian Splattingcited by 3
  5. 2025Iterative Sampling Methods for Sinkhorn Distributionally Robust Optimizationcited by 1