mlsanityData to 7 Oct 2026

POCO: Point Convolution for Surface Reconstruction

arXiv 2201.01831 ↗First posted 5 Jan 2022cs.CV cs.CG cs.LGcited by 91

Not on the map. Why?

Abstract (arXiv)

Implicit neural networks have been successfully used for surface reconstruction from point clouds. However, many of them face scalability issues as they encode the isosurface function of a whole object or scene into a single latent vector. To overcome this limitation, a few approaches infer latent vectors on a coarse regular 3D grid or on 3D patches, and interpolate them to answer occupancy queries. In doing so, they loose the direct connection with the input points sampled on the surface of objects, and they attach information uniformly in space rather than where it matters the most, i.e., near the surface. Besides, relying on fixed patch sizes may require discretization tuning. To address these issues, we propose to use point cloud convolutions and compute latent vectors at each input point. We then perform a learning-based interpolation on nearest neighbors using inferred weights. Experiments on both object and scene datasets show that our approach significantly outperforms other methods on most classical metrics, producing finer details and better reconstructing thinner volumes. The code is available at https://github.com/valeoai/POCO.

Key facts. Cited by 91 papers in the map (data through 7 Oct 2026). Builds on 93 of its 141 references that are in the map. First posted on arXiv 5 Jan 2022.

Builds on

93 of 141 references.

References · 93

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  3. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
  4. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,088
  5. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,880
  6. 2019DeepSDF: Learning Continuous Signed Distance Functions for Shape Representationcited by 2,682
  7. 2020Implicit Neural Representations with Periodic Activation Functionscited by 2,214
  8. 2020Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domainscited by 2,170
  9. 2018Occupancy Networks: Learning 3D Reconstruction in Function Spacecited by 2,037
  10. 20163D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotationcited by 1,471
  11. 2017Matterport3D: Learning from RGB-D Data in Indoor Environmentscited by 1,354
  12. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,202
38 more
  1. 20194D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networkscited by 1,201
  2. 2018Learning Implicit Fields for Generative Shape Modelingcited by 1,060
  3. 2016A Point Set Generation Network for 3D Object Reconstruction from a Single Imagecited by 972
  4. 2020Neural Sparse Voxel Fieldscited by 959
  5. 2019Scene Representation Networks: Continuous 3D-Structure-Aware Neural Scene Representationscited by 864
  6. 2019PIFu: Pixel-Aligned Implicit Function for High-Resolution Clothed Human Digitizationcited by 818
  7. 2018PointCNN: Convolution On X-Transformed Pointscited by 764
  8. 2020Convolutional Occupancy Networkscited by 760
  9. 2020Implicit Geometric Regularization for Learning Shapescited by 707
  10. 2019Differentiable Volumetric Rendering: Learning Implicit 3D Representations without 3D Supervisioncited by 699
  11. 2018PointConv: Deep Convolutional Networks on 3D Point Cloudscited by 656
  12. 2018Pixel2Mesh: Generating 3D Mesh Models from Single RGB Imagescited by 649
  13. 20163D-R2N2: A Unified Approach for Single and Multi-view 3D Object Reconstructioncited by 587
  14. 2016Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modelingcited by 558
  15. 2018AtlasNet: A Papier-Mach'e Approach to Learning 3D Surface Generationcited by 539
  16. 2019PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flowscited by 487
  17. 2021Neural Geometric Level of Detail: Real-time Rendering with Implicit 3D Shapescited by 426
  18. 2019SAL: Sign Agnostic Learning of Shapes from Raw Datacited by 389
  19. 2018PCN: Point Completion Networkcited by 382
  20. 2020Local Implicit Grid Representations for 3D Scenescited by 372
  21. 2019Relation-Shape Convolutional Neural Network for Point Cloud Analysiscited by 346
  22. 2020Implicit Functions in Feature Space for 3D Shape Reconstruction and Completioncited by 343
  23. 2016OctNet: Learning Deep 3D Representations at High Resolutionscited by 340
  24. 2018ABC: A Big CAD Model Dataset For Geometric Deep Learningcited by 339
  25. 2019DISN: Deep Implicit Surface Network for High-quality Single-view 3D Reconstructioncited by 332
  26. 2020Deep Local Shapes: Learning Local SDF Priors for Detailed 3D Reconstructioncited by 321
  27. 2018Learning Category-Specific Mesh Reconstruction from Image Collectionscited by 317
  28. 2020PolyGen: An Autoregressive Generative Model of 3D Meshescited by 267
  29. 2020Neural Unsigned Distance Fields for Implicit Function Learningcited by 265
  30. 2018SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filterscited by 261
  31. 2019What Do Single-view 3D Reconstruction Networks Learn?cited by 259
  32. 2020Learning Gradient Fields for Shape Generationcited by 245
  33. 2019Learning Shape Templates with Structured Implicit Functionscited by 242
  34. 2016Shape Completion using 3D-Encoder-Predictor CNNs and Shape Synthesiscited by 219
  35. 2020MetaSDF: Meta-learning Signed Distance Functionscited by 214
  36. 2019Texture Fields: Learning Texture Representations in Function Spacecited by 210
  37. 2017Octree Generating Networks: Efficient Convolutional Architectures for High-resolution 3D Outputscited by 189
  38. 2021Shape As Points: A Differentiable Poisson Solvercited by 186

50 most cited of 93 All 93 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 18; 2023: 24; 2024: 30; 2025: 14; 2026: 5.

91 papers.

Sort:

By area:

  1. 20233DShape2VecSet: A 3D Shape Representation for Neural Fields and Generative Diffusion Modelscited by 449
  2. 20223D Neural Field Generation using Triplane Diffusioncited by 274
  3. 2023Neural Kernel Surface Reconstructioncited by 102
  4. 2024GSDF: 3DGS Meets SDF for Improved Rendering and Reconstructioncited by 81
  5. 2022ALSO: Automotive Lidar Self-supervision by Occupancy estimationcited by 68
  6. 2022Reconstructing Surfaces for Sparse Point Clouds with On-Surface Priorscited by 60
  7. 2022Surface Reconstruction from Point Clouds by Learning Predictive Context Priorscited by 54
  8. 2024Neural Signed Distance Function Inference through Splatting 3D Gaussians Pulled on Zero-Level Setcited by 47
  9. 2022CAP-UDF: Learning Unsigned Distance Functions Progressively from Raw Point Clouds with Consistency-Aware Field Optimizationcited by 46
  10. 2022Surface Reconstruction from Point Clouds: A Survey and a Benchmarkcited by 41
  11. 2023Towards Better Gradient Consistency for Neural Signed Distance Functions via Level Set Alignmentcited by 39
  12. 2023Learning a More Continuous Zero Level Set in Unsigned Distance Fields through Level Set Projectioncited by 36
38 more
  1. 2023Neural-Singular-Hessian: Implicit Neural Representation of Unoriented Point Clouds by Enforcing Singular Hessiancited by 34
  2. 2022GeoUDF: Surface Reconstruction from 3D Point Clouds via Geometry-guided Distance Representationcited by 28
  3. 2023Unsupervised Inference of Signed Distance Functions from Single Sparse Point Clouds without Learning Priorscited by 28
  4. 2023Learning Signed Distance Functions from Noisy 3D Point Clouds via Noise to Noise Mappingcited by 27
  5. 2022SE(3)-Equivariant Attention Networks for Shape Reconstruction in Function Spacecited by 26
  6. 2023Choose Your Simulator Wisely: A Review on Open-source Simulators for Autonomous Drivingcited by 26
  7. 2022Latent Partition Implicit with Surface Codes for 3D Representationcited by 21
  8. 2022ALTO: Alternating Latent Topologies for Implicit 3D Reconstructioncited by 21
  9. 2023SALUDA: Surface-based Automotive Lidar Unsupervised Domain Adaptationcited by 21
  10. 2024SpaceMesh: A Continuous Representation for Learning Manifold Surface Meshescited by 20
  11. 2024SparseCraft: Few-Shot Neural Reconstruction through Stereopsis Guided Geometric Linearizationcited by 19
  12. 2024Unsupervised Occupancy Learning from Sparse Point Cloudcited by 15
  13. 2024Fast Learning of Signed Distance Functions from Noisy Point Clouds via Noise to Noise Mappingcited by 15
  14. 2022TINC: Tree-structured Implicit Neural Compressioncited by 14
  15. 2025ZeroGrasp: Zero-Shot Shape Reconstruction Enabled Robotic Graspingcited by 13
  16. 2023Mixing-Denoising Generalizable Occupancy Networkscited by 12
  17. 2024NeurCADRecon: Neural Representation for Reconstructing CAD Surfaces by Enforcing Zero Gaussian Curvaturecited by 12
  18. 2022Points2NeRF: Generating Neural Radiance Fields from 3D point cloudcited by 11
  19. 2023GridPull: Towards Scalability in Learning Implicit Representations from 3D Point Cloudscited by 11
  20. 2023A Survey and Benchmark of Automatic Surface Reconstruction from Point Cloudscited by 10
  21. 2023NeuDA: Neural Deformable Anchor for High-Fidelity Implicit Surface Reconstructioncited by 10
  22. 2023Robustifying Generalizable Implicit Shape Networks with a Tunable Non-Parametric Modelcited by 10
  23. 2022What's the Situation with Intelligent Mesh Generation: A Survey and Perspectivescited by 9
  24. 2024Zero-Shot Multi-Object Scene Completioncited by 9
  25. 2022Deep Surface Reconstruction from Point Clouds with Visibility Informationcited by 8
  26. 2022Semi-signed prioritized neural fitting for surface reconstruction from unoriented point cloudscited by 8
  27. 2024DynoSurf: Neural Deformation-based Temporally Consistent Dynamic Surface Reconstructioncited by 8
  28. 2024NeurCross: A Neural Approach to Computing Cross Fields for Quad Mesh Generationcited by 8
  29. 2022OReX: Object Reconstruction from Planar Cross-sections Using Neural Fieldscited by 7
  30. 2023Ins-HOI: Instance Aware Human-Object Interactions Recoverycited by 7
  31. 2024Implicit Filtering for Learning Neural Signed Distance Functions from 3D Point Cloudscited by 7
  32. 2024Inferring Neural Signed Distance Functions by Overfitting on Single Noisy Point Clouds through Finetuning Data-Driven based Priorscited by 7
  33. 2025SparSplat: Fast Multi-View Reconstruction with Generalizable 2D Gaussian Splattingcited by 7
  34. 2025Sparfels: Fast Reconstruction from Sparse Unposed Imagerycited by 7
  35. 2023SuperUDF: Self-supervised UDF Estimation for Surface Reconstructioncited by 6
  36. 2025Parametric Point Cloud Completion for Polygonal Surface Reconstructioncited by 6
  37. 2023Learning Any-View 6DoF Robotic Grasping in Cluttered Scenes via Neural Surface Renderingcited by 5
  38. 2024DDM: A Metric for Comparing 3D Shapes Using Directional Distance Fieldscited by 5

50 most cited of 91 All 91 (JSON)