mlsanityData to 7 Oct 2026

Semantically Stealthy Adversarial Attacks against Segmentation Models

arXiv 2104.01732 ↗First posted 5 Apr 2021cs.CVcited by 7

Not on the map. Why?

Abstract (arXiv)

Segmentation models have been found to be vulnerable to targeted and non-targeted adversarial attacks. However, the resulting segmentation outputs are often so damaged that it is easy to spot an attack. In this paper, we propose semantically stealthy adversarial attacks which can manipulate targeted labels while preserving non-targeted labels at the same time. One challenge is making semantically meaningful manipulations across datasets and models. Another challenge is avoiding damaging non-targeted labels. To solve these challenges, we consider each input image as prior knowledge to generate perturbations. We also design a special regularizer to help extract features. To evaluate our model's performance, we design three basic attack types, namely `vanishing into the context,' `embedding fake labels,' and `displacing target objects.' Our experiments show that our stealthy adversarial model can attack segmentation models with a relatively high success rate on Cityscapes, Mapillary, and BDD100K. Our framework shows good empirical generalization across datasets and models.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 37 of its 48 references that are in the map. First posted on arXiv 5 Apr 2021.

Builds on

37 of 48 references.

References · 37

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  3. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  4. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  5. 2013Intriguing properties of neural networkscited by 5,114
  6. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,845
  7. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
  8. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
  9. 2018Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentationcited by 3,210
  10. 2014Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Imagescited by 774
  11. 2016Universal adversarial perturbationscited by 774
  12. 2017Adversarial Examples for Semantic Segmentation and Object Detectioncited by 275
25 more
  1. 2019Efficient Decision-based Black-box Adversarial Attacks on Face Recognitioncited by 173
  2. 2017Generative Adversarial Perturbationscited by 153
  3. 2017On the Robustness of Semantic Segmentation Models to Adversarial Attackscited by 126
  4. 2020Adversarial Camouflage: Hiding Physical-World Attacks with Natural Stylescited by 126
  5. 2019SemanticAdv: Generating Adversarial Examples via Attribute-conditional Image Editingcited by 93
  6. 2017Universal Adversarial Perturbations Against Semantic Image Segmentationcited by 92
  7. 2019ColorFool: Semantic Adversarial Colorizationcited by 89
  8. 2019AdvPC: Transferable Adversarial Perturbations on 3D Point Cloudscited by 84
  9. 2017Adversarial Attacks Beyond the Image Spacecited by 67
  10. 2020Efficient Adversarial Attacks for Visual Object Trackingcited by 66
  11. 2020Understanding Adversarial Examples from the Mutual Influence of Images and Perturbationscited by 59
  12. 2021A Survey On Universal Adversarial Attackcited by 59
  13. 2019Semantic Adversarial Attacks: Parametric Transformations That Fool Deep Classifierscited by 56
  14. 2017NAG: Network for Adversary Generationcited by 51
  15. 2017Art of singular vectors and universal adversarial perturbationscited by 49
  16. 2020Transferable, Controllable, and Inconspicuous Adversarial Attacks on Person Re-identification With Deep Mis-Rankingcited by 42
  17. 2018Characterizing Adversarial Examples Based on Spatial Consistency Information for Semantic Segmentationcited by 40
  18. 2020CD-UAP: Class Discriminative Universal Adversarial Perturbationcited by 27
  19. 2019Exact Adversarial Attack to Image Captioning via Structured Output Learning with Latent Variablescited by 26
  20. 2021IoU Attack: Towards Temporally Coherent Black-Box Adversarial Attack for Visual Object Trackingcited by 25
  21. 2020Double Targeted Universal Adversarial Perturbationscited by 23
  22. 2020Motion-Excited Sampler: Video Adversarial Attack with Sparked Priorcited by 17
  23. 2018Trust Region Based Adversarial Attack on Neural Networkscited by 11
  24. 2019A Method for Computing Class-wise Universal Adversarial Perturbationscited by 5
  25. 2019Adversarial Examples for Edge Detection: They Exist, and They Transfercited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 2; 2024: 4; 2025: 0; 2026: 1.

7 papers.

  1. 2023Uncertainty-weighted Loss Functions for Improved Adversarial Attacks on Semantic Segmentationcited by 9
  2. 2024Unsegment Anything by Simulating Deformationcited by 6
  3. 2023On Evaluating the Adversarial Robustness of Semantic Segmentation Modelscited by 1
  4. 2024Evaluating the Adversarial Robustness of Semantic Segmentation: Trying Harder Pays Offcited by 1
  5. 2024Detecting Adversarial Attacks in Semantic Segmentation via Uncertainty Estimation: A Deep Analysiscited by 1
  6. 2024Stealthy Multi-Task Adversarial Attackscited by 1
  7. 2026AEGIS: A Semantic GAN and Evidential Learning Frameworkfor Robust Adversarial Detection in Vision Sensorscited by 0