Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
119 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Sample-adaptive Augmentation for Point Cloud Recognition Against Real-world Corruptions (2309.10431v1)

Published 19 Sep 2023 in cs.CV

Abstract: Robust 3D perception under corruption has become an essential task for the realm of 3D vision. While current data augmentation techniques usually perform random transformations on all point cloud objects in an offline way and ignore the structure of the samples, resulting in over-or-under enhancement. In this work, we propose an alternative to make sample-adaptive transformations based on the structure of the sample to cope with potential corruption via an auto-augmentation framework, named as AdaptPoint. Specially, we leverage a imitator, consisting of a Deformation Controller and a Mask Controller, respectively in charge of predicting deformation parameters and producing a per-point mask, based on the intrinsic structural information of the input point cloud, and then conduct corruption simulations on top. Then a discriminator is utilized to prevent the generation of excessive corruption that deviates from the original data distribution. In addition, a perception-guidance feedback mechanism is incorporated to guide the generation of samples with appropriate difficulty level. Furthermore, to address the paucity of real-world corrupted point cloud, we also introduce a new dataset ScanObjectNN-C, that exhibits greater similarity to actual data in real-world environments, especially when contrasted with preceding CAD datasets. Experiments show that our method achieves state-of-the-art results on multiple corruption benchmarks, including ModelNet-C, our ScanObjectNN-C, and ShapeNet-C.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (36)
  1. Shapenet: An information-rich 3d model repository. arXiv preprint arXiv:1512.03012, 2015.
  2. Pointmixup: Augmentation for point clouds. In ECCV, pages 330–345. Springer, 2020.
  3. Imagenet: A large-scale hierarchical image database. In CVPR, pages 248–255. Ieee, 2009.
  4. Poseaug: A differentiable pose augmentation framework for 3d human pose estimation. In CVPR, pages 8575–8584, 2021.
  5. Revisiting point cloud shape classification with a simple and effective baseline. In ICML, pages 3809–3820, 2021.
  6. Benchmarking neural network robustness to common corruptions and perturbations. arXiv preprint arXiv:1903.12261, 2019.
  7. Point cloud augmentation with weighted local transformations. In ICCV, pages 548–557, 2021.
  8. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014.
  9. Regularization strategy for point cloud via rigidly mixed sample. In CVPR, pages 15900–15909, 2021.
  10. Pointaugment: an auto-augmentation framework for point cloud classification. In CVPR, pages 6378–6387, 2020.
  11. Relation-shape convolutional neural network for point cloud analysis. In CVPR, 2019.
  12. Rethinking network design and local geometry in point cloud: A simple residual mlp framework. In ICLR, 2022.
  13. Masked autoencoders for point cloud self-supervised learning. In ECCV, pages 604–621. Springer, 2022.
  14. Pointnet: Deep learning on point sets for 3d classification and segmentation. In CVPR, 2017.
  15. Pointnet++: Deep hierarchical feature learning on point sets in a metric space. In NIPS, 2017.
  16. Pointnext: Revisiting pointnet++ with improved training and scaling strategies. In NIPS, 2022.
  17. Pointcloud-c: Benchmarking and analyzing point cloud perception robustness under corruptions. preprint, 2022.
  18. Benchmarking and analyzing point cloud classification under corruptions. arXiv preprint arXiv:2202.03377, 2022.
  19. Robust adversarial objects against deep learning models. In AAAI, volume 34, pages 954–962, 2020.
  20. Revisiting point cloud classification: A new benchmark dataset and classification model on real-world data. In ICCV, 2019.
  21. Attention is all you need. NIPS, 30, 2017.
  22. Papooling: Graph-based position adaptive aggregation of local geometry in point clouds. arXiv preprint arXiv:2111.14067, 2021.
  23. Dynamic graph cnn for learning on point clouds. TOG, 38(5):1–12, 2019.
  24. Pointconv: Deep convolutional networks on 3d point clouds. In CVPR, 2019.
  25. If-defense: 3d adversarial point cloud defense via implicit function based restoration. arXiv preprint arXiv:2010.05272, 2020.
  26. 3d shapenets: A deep representation for volumetric shapes. In CVPR, 2015.
  27. Generating 3d adversarial point clouds. In CVPR, pages 9136–9144, 2019.
  28. Walk in the cloud: Learning curves for point clouds shape analysis. In ICCV, 2021.
  29. Paconv: Position adaptive convolution with dynamic kernel assembling on point clouds. In CVPR, 2021.
  30. Learning geometry-disentangled representation for complementary understanding of 3d object point cloud. In AAAI, volume 35, pages 3056–3064, 2021.
  31. Adversarial attack and defense on point sets. arXiv preprint arXiv:1902.10899, 2019.
  32. Point-bert: Pre-training 3d point cloud transformers with masked point modeling. In CVPR, pages 19313–19322, 2022.
  33. Point transformer. In ICCV, 2021.
  34. Pointcloud saliency maps. In ICCV, pages 1598–1606, 2019.
  35. Lg-gan: Label guided adversarial network for flexible targeted attack of point cloud based deep networks. In CVPR, pages 10356–10365, 2020.
  36. Dup-net: Denoiser and upsampler network for 3d adversarial point clouds defense. In CVPR, pages 1961–1970, 2019.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (7)
  1. Jie Wang (480 papers)
  2. Lihe Ding (11 papers)
  3. Tingfa Xu (42 papers)
  4. Shaocong Dong (7 papers)
  5. Xinli Xu (17 papers)
  6. Long Bai (87 papers)
  7. Jianan Li (88 papers)
Citations (3)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now