pith. sign in

arxiv: 2412.00404 · v2 · pith:DK326DCMnew · submitted 2024-11-30 · 💻 cs.CV

Hard-Label Black-Box Attacks on 3D Point Clouds

Daizong Liu , Yunbo Tao , Junhao Dong , Keke Tang , Pan Zhou , Wei Hu , Yew-Soon Ong This is my paper

Pith reviewed 2026-05-23 08:34 UTC · model grok-4.3

classification 💻 cs.CV
keywords adversarial attacks3D point cloudsblack-box attackshard-labelspectral fusiondecision boundarypoint cloud classification
0
0 comments X

The pith

A spectrum-aware decision boundary method enables hard-label black-box attacks on 3D point clouds.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper aims to show that adversarial examples for 3D point cloud models can be created when the attacker sees only the final predicted class label. It builds a class-aware decision boundary by fusing point clouds from different classes in the spectral domain with a learnable strategy intended to keep original shapes intact. An iterative optimization then adjusts coordinates and spectra to move samples across that boundary using small changes. This setup is tested against existing attackers that need gradients or logits, claiming better success rates and cleaner adversarial outputs. Readers would care because many deployed 3D systems reveal only labels, so attacks limited to that information are more realistic threats.

Core claim

We introduce a spectrum-aware decision boundary algorithm that first uses a learnable spectrum-fusion strategy to adaptively combine point clouds of different classes in the spectral domain, producing intermediate samples without distorting geometry. An iterative coordinate-spectrum optimization with curvature-aware boundary search then moves these samples along the decision boundary to produce adversarial point clouds with small perturbations.

What carries the argument

Learnable spectrum-fusion strategy inside a spectrum-aware decision boundary algorithm that constructs intermediate samples and searches the boundary with curvature awareness.

If this is right

  • Attacks become feasible in real-world settings where only prediction labels are exposed.
  • Generated adversarial point clouds achieve competitive success rates against white-box and black-box baselines.
  • Adversary quality improves because perturbations remain small and geometry is preserved.
  • The method relies on spectral-domain operations rather than direct coordinate gradient estimates.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar fusion ideas might transfer to hard-label attacks on other 3D tasks such as segmentation if only class outputs are available.
  • Defenses could add spectral-domain checks to detect or block such fused intermediates.
  • The approach might scale to larger scenes if the fusion step is made more efficient.
  • Comparison on datasets with varying point densities would test whether the boundary search remains stable.

Load-bearing premise

Spectral fusion of point clouds from different classes can produce useful intermediate samples near the decision boundary while leaving the original geometry undistorted.

What would settle it

A test set of fused intermediate point clouds whose geometry metrics deviate substantially from the originals, or an optimization run that requires large perturbations to flip labels.

Figures

Figures reproduced from arXiv: 2412.00404 by Daizong Liu, Junhao Dong, Keke Tang, Pan Zhou, Wei Hu, Yew-Soon Ong, Yunbo Tao.

Figure 1
Figure 1. Figure 1: Illustration of our motivation. Our 3D hard-label black-box setting [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overall pipeline of our proposed Hard-Label Black-Box Attack. Specifically, we first design a learnable spectrum-fusion boundary-cloud generation [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Illustration of how we generate the learnable spectrum-fusion rates. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Illustration on our geometry-aware optimization strategy. Specifically, () [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Comparisons on the inference speed and the perturbation size of [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Visualization results of adversarial samples generated by different attack methods. Here, we compare our attack with the SOTA attack methods SI-Adv [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Visualization results of adversarial samples generated by different fusion methods. Here, we list four fusion variants: (1) Learnable fusion with [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Visualization results of adversarial samples generated by different optimization steps. [PITH_FULL_IMAGE:figures/full_fig_p013_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Comparison on the 2D hard-label setting. [PITH_FULL_IMAGE:figures/full_fig_p014_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: 3D attack comparison on more 3D datasets. [PITH_FULL_IMAGE:figures/full_fig_p014_11.png] view at source ↗
read the original abstract

With the maturity of depth sensors in various 3D safety-critical applications, 3D point cloud models have been shown to be vulnerable to adversarial attacks. Almost all existing 3D attackers simply follow the white-box or black-box setting to iteratively update coordinate perturbations based on back-propagated or estimated gradients. However, these methods are hard to deploy in real-world scenarios (no model details are provided) as they severely rely on parameters or output logits of victim models. To this end, we propose point cloud attacks from a more practical setting, i.e., hard-label black-box attack, in which attackers can only access the prediction label of 3D input. We introduce a novel 3D attack method based on a new spectrum-aware decision boundary algorithm to generate high-quality adversarial samples. In particular, we first construct a class-aware model decision boundary, by developing a learnable spectrum-fusion strategy to adaptively fuse point clouds of different classes in the spectral domain, aiming to craft their intermediate samples without distorting the original geometry. Then, we devise an iterative coordinate-spectrum optimization method with curvature-aware boundary search to move the intermediate sample along the decision boundary for generating adversarial point clouds with trivial perturbations. Experiments demonstrate that our attack competitively outperforms existing white/black-box attackers in terms of attack performance and adversary quality.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript proposes a hard-label black-box adversarial attack on 3D point cloud models. It constructs a class-aware decision boundary via a learnable spectrum-fusion strategy that adaptively combines point clouds of different classes in the spectral domain, followed by an iterative coordinate-spectrum optimization using curvature-aware boundary search to produce adversarial examples with minimal perturbations. The central claim is that this method competitively outperforms existing white-box and black-box attackers in both attack success and adversary quality.

Significance. If the experimental claims hold, the work would be significant for adversarial robustness research in 3D vision. Hard-label black-box attacks are more realistic for safety-critical deployments (e.g., autonomous systems) where model internals and logits are unavailable; a spectrum-aware boundary construction could provide new tools for analyzing decision boundaries in point-cloud networks.

major comments (2)
  1. [Abstract] Abstract: the central claim that the method 'competitively outperforms existing white/black-box attackers' is asserted without any quantitative metrics, error bars, dataset names, attack success rates, or ablation results. This prevents verification of the outperformance result that is load-bearing for the paper's contribution.
  2. [Abstract] The description of the learnable spectrum-fusion strategy (abstract) introduces free parameters whose adaptation is claimed to preserve geometry, but no derivation or constraint is shown that guarantees the fused samples remain on the original manifold; this assumption underpins the subsequent curvature-aware search and must be validated.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address each major comment below and have prepared revisions to strengthen the presentation of our contributions.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that the method 'competitively outperforms existing white/black-box attackers' is asserted without any quantitative metrics, error bars, dataset names, attack success rates, or ablation results. This prevents verification of the outperformance result that is load-bearing for the paper's contribution.

    Authors: We agree that the abstract would be strengthened by including concrete quantitative support for the central claim. The full paper reports these details extensively in Sections 4 and 5 (including attack success rates, mean perturbation distances, standard deviations, results on ModelNet40 and ShapeNet, and ablation studies). In the revised manuscript we will expand the abstract with a concise sentence reporting the primary quantitative outcomes to enable immediate verification. revision: yes

  2. Referee: [Abstract] The description of the learnable spectrum-fusion strategy (abstract) introduces free parameters whose adaptation is claimed to preserve geometry, but no derivation or constraint is shown that guarantees the fused samples remain on the original manifold; this assumption underpins the subsequent curvature-aware search and must be validated.

    Authors: The abstract is a high-level summary. The derivation of the spectrum-fusion operator, the learnable parameters, and the explicit constraint (based on the spectral properties of the Fourier basis) that keeps fused samples on the original manifold are presented in Section 3.2 together with the curvature-aware optimization. We will revise the abstract to include a brief reference to this manifold-preservation constraint and add a short empirical validation paragraph in the experiments section. revision: partial

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper proposes an algorithmic construction for hard-label black-box attacks on 3D point clouds, relying on a learnable spectrum-fusion strategy and curvature-aware boundary search. No equations, derivations, or parameter-fitting steps are described that reduce the claimed attack performance or decision boundary construction to inputs defined by the method itself. The central claims rest on the described procedure and experimental validation rather than any self-referential reduction or self-citation chain. This is a standard case of an independent algorithmic contribution with no load-bearing circular steps.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 1 invented entities

Review based on abstract only; the method rests on the unverified assumption that spectral-domain fusion preserves geometry and that the iterative search reliably crosses the decision boundary with trivial perturbations.

free parameters (1)
  • learnable spectrum-fusion parameters
    Parameters that control adaptive fusion of point clouds from different classes; their values are learned during attack generation.
axioms (1)
  • domain assumption Spectral-domain fusion of point clouds from different classes produces intermediate samples that preserve original geometry
    Invoked when constructing the class-aware decision boundary without distortion.
invented entities (1)
  • spectrum-aware decision boundary no independent evidence
    purpose: To enable generation of high-quality adversarial point clouds under hard-label access
    New construct introduced to guide the attack; no independent evidence provided in abstract.

pith-pipeline@v0.9.0 · 5781 in / 1316 out tokens · 38330 ms · 2026-05-23T08:34:20.366425+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

108 extracted references · 108 canonical work pages · 8 internal anchors

  1. [1]

    Intriguing properties of neural networks

    C. Szegedy, W. Zaremba, I. Sutskever, J. Bruna, D. Erhan, I. Goodfel- low, and R. Fergus, “Intriguing properties of neural networks,” arXiv preprint arXiv:1312.6199, 2013

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  2. [2]

    Explaining and harnessing adversarial examples,

    I. J. Goodfellow, J. Shlens, and C. Szegedy, “Explaining and harnessing adversarial examples,” arXiv preprint, 2014

    work page 2014

  3. [3]

    A survey of attacks on large vision- language models: Resources, advances, and future trends

    D. Liu, M. Yang, X. Qu, P. Zhou, Y . Cheng, and W. Hu, “A survey of attacks on large vision-language models: Resources, advances, and future trends,” arXiv preprint arXiv:2407.07403 , 2024

    work page arXiv 2024

  4. [4]

    Pandora’s box: Towards building universal attackers against real-world large vision-language models,

    D. Liu, M. Yang, X. Qu, P. Zhou, X. Fang, K. Tang, Y . Wan, and L. Sun, “Pandora’s box: Towards building universal attackers against real-world large vision-language models,” in The Thirty-eighth Annual Conference on Neural Information Processing Systems , 2024

    work page 2024

  5. [5]

    Boosting adversarial attacks with momentum,

    Y . Dong, F. Liao, T. Pang, H. Su, J. Zhu, X. Hu, and J. Li, “Boosting adversarial attacks with momentum,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2018, pp. 9185–9193

    work page 2018

  6. [6]

    Towards deep learning models resistant to adversarial attacks,

    A. Madry, A. Makelov, L. Schmidt, D. Tsipras, and A. Vladu, “Towards deep learning models resistant to adversarial attacks,” in Proceedings of the International Conference on Learning Representations (ICLR) , 2017

    work page 2017

  7. [7]

    Adversarial Machine Learning at Scale

    A. Kurakin, I. Goodfellow, and S. Bengio, “Adversarial machine learning at scale,” arXiv preprint arXiv:1611.01236 , 2016

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  8. [8]

    Autozoom: Autoencoder-based zeroth order optimiza- tion method for attacking black-box neural networks,

    C.-C. Tu, P. Ting, P.-Y . Chen, S. Liu, H. Zhang, J. Yi, C.-J. Hsieh, and S.-M. Cheng, “Autozoom: Autoencoder-based zeroth order optimiza- tion method for attacking black-box neural networks,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 33, no. 01, 2019, pp. 742–749

    work page 2019

  9. [9]

    Saanet: Siamese action-units attention network for improving dynamic facial expression recognition,

    D. Liu, X. Ouyang, S. Xu, P. Zhou, K. He, and S. Wen, “Saanet: Siamese action-units attention network for improving dynamic facial expression recognition,” Neurocomputing, vol. 413, pp. 145–157, 2020

    work page 2020

  10. [10]

    Spatiotemporal graph neural network based mask reconstruction for video object segmentation,

    D. Liu, S. Xu, X.-Y . Liu, Z. Xu, W. Wei, and P. Zhou, “Spatiotemporal graph neural network based mask reconstruction for video object segmentation,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 35, no. 3, 2021, pp. 2100–2108

    work page 2021

  11. [11]

    Few-shot tempo- ral sentence grounding via memory-guided semantic learning,

    D. Liu, P. Zhou, Z. Xu, H. Wang, and R. Li, “Few-shot tempo- ral sentence grounding via memory-guided semantic learning,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 33, no. 5, pp. 2491–2505, 2022

    work page 2022

  12. [12]

    Multi-view 3d object detection network for autonomous driving,

    X. Chen, H. Ma, J. Wan, B. Li, and T. Xia, “Multi-view 3d object detection network for autonomous driving,” in Proceedings of the IEEE conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 1907–1915

    work page 2017

  13. [13]

    Eye in the sky: Drone- based object tracking and 3d localization,

    H. Zhang, G. Wang, Z. Lei, and J.-N. Hwang, “Eye in the sky: Drone- based object tracking and 3d localization,” in Proceedings of the 27th ACM international conference on multimedia , 2019, pp. 899–907

    work page 2019

  14. [14]

    Density-insensitive unsupervised domain adaption on 3d object detection,

    Q. Hu, D. Liu, and W. Hu, “Density-insensitive unsupervised domain adaption on 3d object detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 17 556–17 566

    work page 2023

  15. [15]

    Shape completion enabled robotic grasping,

    J. Varley, C. DeChant, A. Richardson, J. Ruales, and P. Allen, “Shape completion enabled robotic grasping,” in IEEE international Confer- ence on Intelligent Robots and Systems (IROS) , 2017, pp. 2442–2447

    work page 2017

  16. [16]

    Reliable vision-based grasping target recognition for upper limb prostheses,

    B. Zhong, H. Huang, and E. Lobaton, “Reliable vision-based grasping target recognition for upper limb prostheses,” IEEE Transactions on Cybernetics, 2020

    work page 2020

  17. [17]

    Dense object grounding in 3d scenes,

    W. Huang, D. Liu, and W. Hu, “Dense object grounding in 3d scenes,” in Proceedings of the 31st ACM International Conference on Multimedia, 2023, pp. 5017–5026

    work page 2023

  18. [18]

    A survey on text-guided 3d visual grounding: Elements, recent advances, and future directions,

    D. Liu, Y . Liu, W. Huang, and W. Hu, “A survey on text-guided 3d visual grounding: Elements, recent advances, and future directions,” arXiv preprint arXiv:2406.05785 , 2024

    work page arXiv 2024

  19. [19]

    Cross-task knowledge transfer for semi-supervised joint 3d grounding and captioning,

    Y . Liu, D. Liu, Z. Guo, and W. Hu, “Cross-task knowledge transfer for semi-supervised joint 3d grounding and captioning,” in Proceedings of the 32nd ACM International Conference on Multimedia , 2024, pp. 3818–3827

    work page 2024

  20. [20]

    Advancing 3d object grounding beyond a single 3d scene,

    W. Huang, D. Liu, and W. Hu, “Advancing 3d object grounding beyond a single 3d scene,” in Proceedings of the 32nd ACM International Conference on Multimedia , 2024, pp. 7995–8004

    work page 2024

  21. [21]

    Joint top-down and bottom-up frameworks for 3d visual grounding,

    Y . Liu, D. Liu, and W. Hu, “Joint top-down and bottom-up frameworks for 3d visual grounding,” arXiv preprint arXiv:2410.15615 , 2024

    work page arXiv 2024

  22. [22]

    3d deep learning on medical images: a review,

    S. P. Singh, L. Wang, S. Gupta, H. Goli, P. Padmanabhan, and B. Guly ´as, “3d deep learning on medical images: a review,” Sensors, vol. 20, no. 18, p. 5097, 2020

    work page 2020

  23. [23]

    Robust adversarial objects against deep learning models,

    T. Tsai, K. Yang, T.-Y . Ho, and Y . Jin, “Robust adversarial objects against deep learning models,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 34, no. 01, 2020, pp. 954–962

    work page 2020

  24. [24]

    On isometry robustness of deep 3d point cloud models under adversarial attacks,

    Y . Zhao, Y . Wu, C. Chen, and A. Lim, “On isometry robustness of deep 3d point cloud models under adversarial attacks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020, pp. 1201–1210

    work page 2020

  25. [25]

    Lg-gan: Label guided adversarial network for flex- ible targeted attack of point cloud based deep networks,

    H. Zhou, D. Chen, J. Liao, K. Chen, X. Dong, K. Liu, W. Zhang, G. Hua, and N. Yu, “Lg-gan: Label guided adversarial network for flex- ible targeted attack of point cloud based deep networks,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020, pp. 10 356–10 365

    work page 2020

  26. [26]

    Geometry-aware generation of adversarial point clouds,

    Y . Wen, J. Lin, K. Chen, C. P. Chen, and K. Jia, “Geometry-aware generation of adversarial point clouds,” IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) , 2020

    work page 2020

  27. [27]

    Robust geometry-dependent attack for 3d point clouds,

    D. Liu, W. Hu, and X. Li, “Robust geometry-dependent attack for 3d point clouds,” IEEE Transactions on Multimedia , 2023

    work page 2023

  28. [28]

    Explicitly perceiving and preserving the local geometric structures for 3d point cloud attack,

    D. Liu and W. Hu, “Explicitly perceiving and preserving the local geometric structures for 3d point cloud attack,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 38, no. 4, 2024, pp. 3576–3584

    work page 2024

  29. [29]

    Frequency-aware gan for imperceptible transfer attack on 3d point clouds,

    X. Cai, Y . Tao, D. Liu, P. Zhou, X. Qu, J. Dong, K. Tang, and L. Sun, “Frequency-aware gan for imperceptible transfer attack on 3d point clouds,” in Proceedings of the 32nd ACM International Conference on Multimedia, 2024, pp. 6162–6171. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 15

    work page 2024

  30. [30]

    Hiding imperceptible noise in curvature-aware patches for 3d point cloud attack,

    M. Yang, D. Liu, K. Tang, P. Zhou, L. Chen, and J. Chen, “Hiding imperceptible noise in curvature-aware patches for 3d point cloud attack,” in European Conference on Computer Vision. Springer, 2025, pp. 431–448

    work page 2025

  31. [31]

    Adversarial attack and defense on point sets,

    Q. Zhang, J. Yang, R. Fang, B. Ni, J. Liu, and Q. Tian, “Adversarial attack and defense on point sets,” arXiv preprint, 2019

    work page 2019

  32. [32]

    Generating 3d adversarial point clouds,

    C. Xiang, C. R. Qi, and B. Li, “Generating 3d adversarial point clouds,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019, pp. 9136–9144

    work page 2019

  33. [33]

    Shape- invariant 3d adversarial point clouds,

    Q. Huang, X. Dong, D. Chen, H. Zhou, W. Zhang, and N. Yu, “Shape- invariant 3d adversarial point clouds,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2022, pp. 15 335–15 344

    work page 2022

  34. [34]

    Qeba: Query-efficient boundary-based blackbox attack,

    H. Li, X. Xu, X. Zhang, S. Yang, and B. Li, “Qeba: Query-efficient boundary-based blackbox attack,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2020, pp. 1221–1230

    work page 2020

  35. [35]

    Decision-based adver- sarial attack with frequency mixup,

    X.-C. Li, X.-Y . Zhang, F. Yin, and C.-L. Liu, “Decision-based adver- sarial attack with frequency mixup,” IEEE Transactions on Information Forensics and Security, vol. 17, pp. 1038–1052, 2022

    work page 2022

  36. [36]

    Nonlinear pro- jection based gradient estimation for query efficient blackbox attacks,

    H. Li, L. Li, X. Xu, X. Zhang, S. Yang, and B. Li, “Nonlinear pro- jection based gradient estimation for query efficient blackbox attacks,” in International Conference on Artificial Intelligence and Statistics . PMLR, 2021, pp. 3142–3150

    work page 2021

  37. [37]

    Black-box decision based adversarial attack with symmetric α- stable distribution,

    V . Srinivasan, E. E. Kuruoglu, K.-R. M ¨uller, W. Samek, and S. Naka- jima, “Black-box decision based adversarial attack with symmetric α- stable distribution,” in 2019 27th European Signal Processing Confer- ence (EUSIPCO). IEEE, 2019, pp. 1–5

    work page 2019

  38. [38]

    Black-box adversarial attacks with limited queries and information,

    A. Ilyas, L. Engstrom, A. Athalye, and J. Lin, “Black-box adversarial attacks with limited queries and information,” in International Confer- ence on Machine Learning . PMLR, 2018, pp. 2137–2146

    work page 2018

  39. [39]

    Large-capacity image steganography based on invertible neural networks,

    S.-P. Lu, R. Wang, T. Zhong, and P. L. Rosin, “Large-capacity image steganography based on invertible neural networks,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2021, pp. 10 816–10 825

    work page 2021

  40. [40]

    Robust invertible image steganography,

    Y . Xu, C. Mou, Y . Hu, J. Xie, and J. Zhang, “Robust invertible image steganography,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2022, pp. 7875–7884

    work page 2022

  41. [41]

    Graph signal processing for geometric data and beyond: Theory and applications,

    W. Hu, J. Pang, X. Liu, D. Tian, C.-W. Lin, and A. Vetro, “Graph signal processing for geometric data and beyond: Theory and applications,” IEEE Transactions on Multimedia , 2021

    work page 2021

  42. [42]

    Exploring the devil in graph spectral domain for 3d point cloud attacks,

    Q. Hu, D. Liu, and W. Hu, “Exploring the devil in graph spectral domain for 3d point cloud attacks,” arXiv preprint arXiv:2202.07261 , 2022

    work page arXiv 2022

  43. [43]

    The emerging field of signal processing on graphs: Ex- tending high-dimensional data analysis to networks and other irregular domains,

    D. I. Shuman, S. K. Narang, P. Frossard, A. Ortega, and P. Van- dergheynst, “The emerging field of signal processing on graphs: Ex- tending high-dimensional data analysis to networks and other irregular domains,” IEEE Signal Process. Mag., vol. 30, no. 3, pp. 83–98, 2013

    work page 2013

  44. [44]

    Graph signal processing: Overview, challenges, and ap- plications,

    A. Ortega, P. Frossard, J. Kova ˇcevi´c, J. M. Moura, and P. Van- dergheynst, “Graph signal processing: Overview, challenges, and ap- plications,” Proceedings of the IEEE , vol. 106, no. 5, pp. 808–828, 2018

    work page 2018

  45. [45]

    Graph spectral image processing,

    G. Cheung, E. Magli, Y . Tanaka, and M. K. Ng, “Graph spectral image processing,” Proceedings of the IEEE , vol. 106, no. 5, pp. 907–930, 2018

    work page 2018

  46. [46]

    Ortega, Introduction to graph signal processing

    A. Ortega, Introduction to graph signal processing . Cambridge University Press, 2022

    work page 2022

  47. [47]

    Wavelets on graphs via spectral graph theory,

    D. K. Hammond, P. Vandergheynst, and R. Gribonval, “Wavelets on graphs via spectral graph theory,” Appl. Comput. Harmonic Anal. , vol. 30, no. 2, pp. 129–150, 2011

    work page 2011

  48. [48]

    3dhacker: Spectrum-based decision boundary generation for hard-label 3d point cloud attack,

    Y . Tao, D. Liu, P. Zhou, Y . Xie, W. Du, and W. Hu, “3dhacker: Spectrum-based decision boundary generation for hard-label 3d point cloud attack,” in Proceedings of the IEEE International Conference on Computer Vision (ICCV) , 2023

    work page 2023

  49. [49]

    Multi-view convolutional neural networks for 3d shape recognition,

    H. Su, S. Maji, E. Kalogerakis, and E. Learned-Miller, “Multi-view convolutional neural networks for 3d shape recognition,” in Proceed- ings of the IEEE International Conference on Computer Vision (ICCV), 2015, pp. 945–953

    work page 2015

  50. [50]

    Multi-view harmonized bilinear network for 3d object recognition,

    T. Yu, J. Meng, and J. Yuan, “Multi-view harmonized bilinear network for 3d object recognition,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 186–194

    work page 2018

  51. [51]

    Pointcnn: Convolution on x-transformed points,

    Y . Li, R. Bu, M. Sun, W. Wu, X. Di, and B. Chen, “Pointcnn: Convolution on x-transformed points,” Advances in Neural Information Processing Systems (NIPS) , vol. 31, pp. 820–830, 2018

    work page 2018

  52. [52]

    Deep sets,

    M. Zaheer, S. Kottur, S. Ravanbakhsh, B. Poczos, R. R. Salakhutdi- nov, and A. J. Smola, “Deep sets,” Advances in Neural Information Processing Systems (NIPS) , vol. 30, 2017

    work page 2017

  53. [53]

    Pointnet: Deep learning on point sets for 3d classification and segmentation,

    C. R. Qi, H. Su, K. Mo, and L. J. Guibas, “Pointnet: Deep learning on point sets for 3d classification and segmentation,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 652–660

    work page 2017

  54. [54]

    Pointnet++: Deep hierarchical feature learning on point sets in a metric space,

    C. R. Qi, L. Yi, H. Su, and L. J. Guibas, “Pointnet++: Deep hierarchical feature learning on point sets in a metric space,” Advances in Neural Information Processing Systems (NIPS) , 2017

    work page 2017

  55. [55]

    Structural relational reasoning of point clouds,

    Y . Duan, Y . Zheng, J. Lu, J. Zhou, and Q. Tian, “Structural relational reasoning of point clouds,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019, pp. 949–958

    work page 2019

  56. [56]

    Densepoint: Learning densely contextual representation for efficient point cloud processing,

    Y . Liu, B. Fan, G. Meng, J. Lu, S. Xiang, and C. Pan, “Densepoint: Learning densely contextual representation for efficient point cloud processing,” in Proceedings of the IEEE International Conference on Computer Vision (ICCV) , 2019, pp. 5239–5248

    work page 2019

  57. [57]

    Mod- eling point clouds with self-attention and gumbel subset sampling,

    J. Yang, Q. Zhang, B. Ni, L. Li, J. Liu, M. Zhou, and Q. Tian, “Mod- eling point clouds with self-attention and gumbel subset sampling,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019, pp. 3323–3332

    work page 2019

  58. [58]

    Paconv: Position adaptive convolution with dynamic kernel assembling on point clouds,

    M. Xu, R. Ding, H. Zhao, and X. Qi, “Paconv: Position adaptive convolution with dynamic kernel assembling on point clouds,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 3173–3182

    work page 2021

  59. [59]

    Point Convolutional Neural Networks by Extension Operators

    M. Atzmon, H. Maron, and Y . Lipman, “Point convolutional neural networks by extension operators,” arXiv preprint arXiv:1803.10091 , 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  60. [60]

    Kpconv: Flexible and deformable convolution for point clouds,

    H. Thomas, C. R. Qi, J.-E. Deschaud, B. Marcotegui, F. Goulette, and L. J. Guibas, “Kpconv: Flexible and deformable convolution for point clouds,” in Proceedings of the IEEE International Conference on Computer Vision (ICCV) , 2019, pp. 6411–6420

    work page 2019

  61. [61]

    Relation-shape convolutional neural network for point cloud analysis,

    Y . Liu, B. Fan, S. Xiang, and C. Pan, “Relation-shape convolutional neural network for point cloud analysis,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2019, pp. 8895–8904

    work page 2019

  62. [62]

    Dynamic edge-conditioned fil- ters in convolutional neural networks on graphs,

    M. Simonovsky and N. Komodakis, “Dynamic edge-conditioned fil- ters in convolutional neural networks on graphs,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 3693–3702

    work page 2017

  63. [63]

    Mining point cloud local structures by kernel correlation and graph pooling,

    Y . Shen, C. Feng, Y . Yang, and D. Tian, “Mining point cloud local structures by kernel correlation and graph pooling,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 4548–4557

    work page 2018

  64. [64]

    Dynamic graph cnn for learning on point clouds,

    Y . Wang, Y . Sun, Z. Liu, S. E. Sarma, M. M. Bronstein, and J. M. Solomon, “Dynamic graph cnn for learning on point clouds,” ACM Transactions on Graphics (tog) , vol. 38, no. 5, pp. 1–12, 2019

    work page 2019

  65. [65]

    Grid-gcn for fast and scalable point cloud learning,

    Q. Xu, X. Sun, C.-Y . Wu, P. Wang, and U. Neumann, “Grid-gcn for fast and scalable point cloud learning,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2020, pp. 5661–5670

    work page 2020

  66. [66]

    Graphter: Unsupervised learning of graph transformation equivariant representations via auto-encoding node-wise transformations,

    X. Gao, W. Hu, and G.-J. Qi, “Graphter: Unsupervised learning of graph transformation equivariant representations via auto-encoding node-wise transformations,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2020, pp. 7163– 7172

    work page 2020

  67. [67]

    Rgcnn: Regularized graph cnn for point cloud segmentation,

    G. Te, W. Hu, A. Zheng, and Z. Guo, “Rgcnn: Regularized graph cnn for point cloud segmentation,” in Proceedings of the 26th ACM international conference on Multimedia , 2018, pp. 746–754

    work page 2018

  68. [68]

    Self-contrastive learning with hard negative sampling for self-supervised point cloud learning,

    B. Du, X. Gao, W. Hu, and X. Li, “Self-contrastive learning with hard negative sampling for self-supervised point cloud learning,” in Proceedings of the 29th ACM International Conference on Multimedia, 2021, pp. 3133–3142

    work page 2021

  69. [69]

    Deepfool: a sim- ple and accurate method to fool deep neural networks,

    S.-M. Moosavi-Dezfooli, A. Fawzi, and P. Frossard, “Deepfool: a sim- ple and accurate method to fool deep neural networks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 2574–2582

    work page 2016

  70. [70]

    Univer- sal adversarial perturbations,

    S.-M. Moosavi-Dezfooli, A. Fawzi, O. Fawzi, and P. Frossard, “Univer- sal adversarial perturbations,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 1765– 1773

    work page 2017

  71. [71]

    Robustness of 3d deep learning in an adversarial setting,

    M. Wicker and M. Kwiatkowska, “Robustness of 3d deep learning in an adversarial setting,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2019, pp. 11 767– 11 775

    work page 2019

  72. [72]

    Pointcloud saliency maps,

    T. Zheng, C. Chen, J. Yuan, B. Li, and K. Ren, “Pointcloud saliency maps,” in Proceedings of the IEEE International Conference on Com- puter Vision (ICCV) , 2019, pp. 1598–1606. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 16

    work page 2019

  73. [73]

    Pct: Point cloud transformer,

    M.-H. Guo, J.-X. Cai, Z.-N. Liu, T.-J. Mu, R. R. Martin, and S.-M. Hu, “Pct: Point cloud transformer,”Computational Visual Media, vol. 7, pp. 187–199, 2021

    work page 2021

  74. [74]

    Point transformer,

    H. Zhao, L. Jiang, J. Jia, P. H. Torr, and V . Koltun, “Point transformer,” in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 16 259–16 268

    work page 2021

  75. [75]

    Point-bert: Pre-training 3d point cloud transformers with masked point modeling,

    X. Yu, L. Tang, Y . Rao, T. Huang, J. Zhou, and J. Lu, “Point-bert: Pre-training 3d point cloud transformers with masked point modeling,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 19 313–19 322

    work page 2022

  76. [76]

    Pvt: Point-voxel transformer for point cloud learning,

    C. Zhang, H. Wan, X. Shen, and Z. Wu, “Pvt: Point-voxel transformer for point cloud learning,” International Journal of Intelligent Systems , vol. 37, no. 12, pp. 11 985–12 008, 2022

    work page 2022

  77. [77]

    Re- visiting point cloud classification: A new benchmark dataset and clas- sification model on real-world data,

    M. A. Uy, Q.-H. Pham, B.-S. Hua, T. Nguyen, and S.-K. Yeung, “Re- visiting point cloud classification: A new benchmark dataset and clas- sification model on real-world data,” in Proceedings of the IEEE/CVF international conference on computer vision , 2019, pp. 1588–1597

    work page 2019

  78. [78]

    ShapeNet: An Information-Rich 3D Model Repository

    A. X. Chang, T. Funkhouser, L. Guibas, P. Hanrahan, Q. Huang, Z. Li, S. Savarese, M. Savva, S. Song, H. Su et al. , “Shapenet: An information-rich 3d model repository,” arXiv preprint arXiv:1512.03012, 2015

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  79. [79]

    3d shapenets: A deep representation for volumetric shapes,

    Z. Wu, S. Song, A. Khosla, F. Yu, L. Zhang, X. Tang, and J. Xiao, “3d shapenets: A deep representation for volumetric shapes,” in Pro- ceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 1912–1920

    work page 2015

  80. [80]

    Extending adversarial attacks and defenses to deep 3d point cloud classifiers,

    D. Liu, R. Yu, and H. Su, “Extending adversarial attacks and defenses to deep 3d point cloud classifiers,” in 2019 IEEE International Con- ference on Image Processing (ICIP) , 2019, pp. 2279–2283

    work page 2019

Showing first 80 references.