mlsanityData to 7 Oct 2026

Ada3Diff: Defending against 3D Adversarial Point Clouds via Adaptive Diffusion

arXiv 2211.16247 ↗First posted 29 Nov 2022cs.CVcited by 8

Not on the map. Why?

Abstract (arXiv)

Deep 3D point cloud models are sensitive to adversarial attacks, which poses threats to safety-critical applications such as autonomous driving. Robust training and defend-by-denoising are typical strategies for defending adversarial perturbations. However, they either induce massive computational overhead or rely heavily upon specified priors, limiting generalized robustness against attacks of all kinds. To remedy it, this paper introduces a novel distortion-aware defense framework that can rebuild the pristine data distribution with a tailored intensity estimator and a diffusion model. To perform distortion-aware forward diffusion, we design a distortion estimation algorithm that is obtained by summing the distance of each point to the best-fitting plane of its local neighboring points, which is based on the observation of the local spatial properties of the adversarial point cloud. By iterative diffusion and reverse denoising, the perturbed point cloud under various distortions can be restored back to a clean distribution. This approach enables effective defense against adaptive attacks with varying noise budgets, enhancing the robustness of existing 3D deep recognition models.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 29 of its 46 references that are in the map. First posted on arXiv 29 Nov 2022.

Builds on

29 of 46 references.

References · 29

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  3. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  4. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  5. 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,220
  6. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,259
  7. 2016A Point Set Generation Network for 3D Object Reconstruction from a Single Imagecited by 972
  8. 2020Convolutional Occupancy Networkscited by 760
  9. 2021Diffusion Probabilistic Models for 3D Point Cloud Generationcited by 734
  10. 2020PCT: Point cloud transformercited by 730
  11. 2018PointConv: Deep Convolutional Networks on 3D Point Cloudscited by 656
  12. 2022LION: Latent Point Diffusion Models for 3D Shape Generationcited by 552
17 more
  1. 20213D Shape Generation and Completion through Point-Voxel Diffusioncited by 516
  2. 2022Diffusion Models for Adversarial Purificationcited by 469
  3. 2019Point-Voxel CNN for Efficient 3D Deep Learningcited by 343
  4. 2021Walk in the Cloud: Learning Curves for Point Clouds Shape Analysiscited by 172
  5. 2018Generating 3D Adversarial Point Cloudscited by 144
  6. 2018PointCloud Saliency Mapscited by 112
  7. 2019Extending Adversarial Attacks and Defenses to Deep 3D Point Cloud Classifierscited by 92
  8. 2022Guided Diffusion Model for Adversarial Purificationcited by 90
  9. 2019AdvPC: Transferable Adversarial Perturbations on 3D Point Cloudscited by 84
  10. 2019Adversarial Attack and Defense on Point Setscited by 63
  11. 2020LG-GAN: Label Guided Adversarial Network for Flexible Targeted Attack of Point Cloud-based Deep Networkscited by 55
  12. 2022Shape-invariant 3D Adversarial Point Cloudscited by 49
  13. 2020IF-Defense: 3D Adversarial Point Cloud Defense via Implicit Function based Restorationcited by 46
  14. 2020Minimal Adversarial Examples for Deep Learning on 3D Point Cloudscited by 41
  15. 2022Boosting 3D Adversarial Attacks with Attacking On Frequencycited by 18
  16. 2022Robust Structured Declarative Classifiers for 3D Point Clouds: Defending Adversarial Attacks with Implicit Gradientscited by 17
  17. 2022PointDP: Diffusion-driven Purification against Adversarial Attacks on 3D Point Cloud Recognitioncited by 16

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 3; 2024: 2; 2025: 2; 2026: 1.

8 papers.

  1. 2023Diffusion-Based Adversarial Sample Generation for Improved Stealthiness and Controllabilitycited by 66
  2. 2023DiffAttack: Evasion Attacks Against Diffusion-Based Adversarial Purificationcited by 36
  3. 2023Adversarial Attacks and Defenses on 3D Point Cloud Classification: A Surveycited by 5
  4. 2024Robust 3D Point Clouds Classification based on Declarative Defenderscited by 2
  5. 2024Enhancing Robustness to Noise Corruption for Point Cloud Recognition via Spatial Sorting and Set-Mixing Aggregation Modulecited by 0
  6. 2025KNN-Defense: Defense against 3D Adversarial Point Clouds using Nearest-Neighbor Searchcited by 0
  7. 2025Multimodal Robust Prompt Distillation for 3D Point Cloud Modelscited by 0
  8. 2026PWAVEP: Purifying Imperceptible Adversarial Perturbations in 3D Point Clouds via Spectral Graph Waveletscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Generating Adversarial Point Clouds Using Diffusion Model
  2. 2024Transferable 3D Adversarial Shape Completion using Diffusion Models
  3. 2026Meltdown: Circuits and Bifurcations in Point-Cloud-Conditioned 3D Diffusion Transformers
  4. 2021Local Aggressive Adversarial Attacks on 3D Point Cloud
  5. 2023Fast Training of Diffusion Transformer with Extreme Masking for 3D Point Clouds Generation
  6. 2024Learning Robust 3D Representation from CLIP via Dual Denoising
  7. 2026Adversarially Guided Diffusion for LiDAR Range Image Synthesis
  8. 2023PC2: Projection-Conditioned Point Cloud Diffusion for Single-Image 3D Reconstruction