mlsanityData to 7 Oct 2026

LightSAL: Lightweight Sign Agnostic Learning for Implicit Surface Representation

arXiv 2103.14273 ↗First posted 26 Mar 2021cs.CVcited by 7

Not on the map. Why?

Abstract (arXiv)

Recently, several works have addressed modeling of 3D shapes using deep neural networks to learn implicit surface representations. Up to now, the majority of works have concentrated on reconstruction quality, paying little or no attention to model size or training time. This work proposes LightSAL, a novel deep convolutional architecture for learning 3D shapes; the proposed work concentrates on efficiency both in network training time and resulting model size. We build on the recent concept of Sign Agnostic Learning for training the proposed network, relying on signed distance fields, with unsigned distance as ground truth. In the experimental section of the paper, we demonstrate that the proposed architecture outperforms previous work in model size and number of required training iterations, while achieving equivalent accuracy. Experiments are based on the D-Faust dataset that contains 41k 3D scans of human shapes. The proposed model has been implemented in PyTorch.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 43 of its 61 references that are in the map. First posted on arXiv 26 Mar 2021.

Builds on

43 of 61 references.

References · 43

  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. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
  5. 2019DeepSDF: Learning Continuous Signed Distance Functions for Shape Representationcited by 2,682
  6. 2020Implicit Neural Representations with Periodic Activation Functionscited by 2,214
  7. 2018Occupancy Networks: Learning 3D Reconstruction in Function Spacecited by 2,037
  8. 2017Deep Setscited by 1,394
  9. 2018Learning Implicit Fields for Generative Shape Modelingcited by 1,060
  10. 2016Geometric deep learning: going beyond Euclidean datacited by 1,042
  11. 2016A Point Set Generation Network for 3D Object Reconstruction from a Single Imagecited by 972
  12. 2019PIFu: Pixel-Aligned Implicit Function for High-Resolution Clothed Human Digitizationcited by 818
31 more
  1. 2020Convolutional Occupancy Networkscited by 760
  2. 2018Pixel2Mesh: Generating 3D Mesh Models from Single RGB Imagescited by 649
  3. 20163D-R2N2: A Unified Approach for Single and Multi-view 3D Object Reconstructioncited by 587
  4. 2020Learning Continuous Image Representation with Local Implicit Image Functioncited by 550
  5. 2018AtlasNet: A Papier-Mach'e Approach to Learning 3D Surface Generationcited by 539
  6. 2019PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flowscited by 487
  7. 2019SAL: Sign Agnostic Learning of Shapes from Raw Datacited by 389
  8. 2020Local Implicit Grid Representations for 3D Scenescited by 372
  9. 2018Learning Category-Specific Mesh Reconstruction from Image Collectionscited by 317
  10. 2018Generating 3D faces using Convolutional Mesh Autoencoderscited by 267
  11. 2020Neural Unsigned Distance Fields for Implicit Function Learningcited by 265
  12. 2016Shape Completion using 3D-Encoder-Predictor CNNs and Shape Synthesiscited by 219
  13. 2019Texture Fields: Learning Texture Representations in Function Spacecited by 210
  14. 2017Octree Generating Networks: Efficient Convolutional Architectures for High-resolution 3D Outputscited by 189
  15. 2019DeepHuman: 3D Human Reconstruction from a Single Imagecited by 182
  16. 2017Multi-view Supervision for Single-view Reconstruction via Differentiable Ray Consistencycited by 177
  17. 2019NASA: Neural Articulated Shape Approximationcited by 149
  18. 2018BodyNet: Volumetric Inference of 3D Human Body Shapescited by 141
  19. 2017Learning Efficient Point Cloud Generation for Dense 3D Object Reconstructioncited by 133
  20. 2017SurfaceNet: An End-to-end 3D Neural Network for Multiview Stereopsiscited by 128
  21. 2020SALD: Sign Agnostic Learning with Derivativescited by 115
  22. 2019Efficient Learning on Point Clouds with Basis Point Setscited by 101
  23. 2017Hierarchical Surface Prediction for 3D Object Reconstructioncited by 89
  24. 2016Unsupervised Learning of 3D Structure from Imagescited by 67
  25. 2019Deep Level Sets: Implicit Surface Representations for 3D Shape Inferencecited by 52
  26. 2017OctNetFusion: Learning Depth Fusion from Datacited by 45
  27. 2020Implicit Feature Networks for Texture Completion from Partial 3D Datacited by 31
  28. 2019Photometric Mesh Optimization for Video-Aligned 3D Object Reconstructioncited by 28
  29. 2019Local Deep Implicit Functions for 3D Shapecited by 18
  30. 2018Volumetric performance capture from minimal camera viewpointscited by 17
  31. 20173D Shape Segmentation via Shape Fully Convolutional Networkscited by 15

Built on by

Cited by year

Citing papers per year: 2021: 2; 2022: 4; 2023: 1.

7 papers.

  1. 2021Beyond Periodicity: Towards a Unifying Framework for Activations in Coordinate-MLPscited by 134
  2. 2022Surface Reconstruction from Point Clouds: A Survey and a Benchmarkcited by 41
  3. 2021Learning Positional Embeddings for Coordinate-MLPscited by 7
  4. 2022On Regularizing Coordinate-MLPscited by 5
  5. 2022How You Start Matters for Generalizationcited by 5
  6. 2023Optimizing Implicit Neural Representations from Point Clouds via Energy-Based Modelscited by 1
  7. 2022Convolutional Neural Network-based Efficient Dense Point Cloud Generation using Unsigned Distance Fieldscited by 0