mlsanityData to 7 Oct 2026

Erosion Attack for Adversarial Training to Enhance Semantic Segmentation Robustness

arXiv 2601.14950 ↗First posted 21 Jan 2026cs.CVcited by 3

Not on the map. Why?

Abstract (arXiv)

Existing segmentation models exhibit significant vulnerability to adversarial attacks.To improve robustness, adversarial training incorporates adversarial examples into model training. However, existing attack methods consider only global semantic information and ignore contextual semantic relationships within the samples, limiting the effectiveness of adversarial training. To address this issue, we propose EroSeg-AT, a vulnerability-aware adversarial training framework that leverages EroSeg to generate adversarial examples. EroSeg first selects sensitive pixels based on pixel-level confidence and then progressively propagates perturbations to higher-confidence pixels, effectively disrupting the semantic consistency of the samples. Experimental results show that, compared to existing methods, our approach significantly improves attack effectiveness and enhances model robustness under adversarial training.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 24 of its 30 references that are in the map. First posted on arXiv 21 Jan 2026.

Builds on

24 of 30 references.

References · 24

  1. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  2. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  3. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
  4. 2016DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFscited by 3,239
  5. 2016Pyramid Scene Parsing Networkcited by 2,472
  6. 2014Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFscited by 552
  7. 2017Adversarial Examples for Semantic Segmentation and Object Detectioncited by 275
  8. 2023AdvCLIP: Downstream-agnostic Adversarial Examples in Multimodal Contrastive Learningcited by 101
  9. 2018CGNet: A Light-weight Context Guided Network for Semantic Segmentationcited by 94
  10. 2022SegPGD: An Effective and Efficient Adversarial Attack for Evaluating and Boosting Segmentation Robustnesscited by 74
  11. 2023MSeg3D: Multi-modal 3D Semantic Segmentation for Autonomous Drivingcited by 68
  12. 2020Dynamic Divide-and-Conquer Adversarial Training for Robust Semantic Segmentationcited by 38
12 more
  1. 2023Downstream-agnostic Adversarial Examplescited by 32
  2. 2024DarkSAM: Fooling Segment Anything Model to Segment Nothingcited by 32
  3. 2023CosPGD: an efficient white-box adversarial attack for pixel-wise prediction taskscited by 27
  4. 2024Transferable Adversarial Facial Images for Privacy Protectioncited by 26
  5. 2024Securely Fine-tuning Pre-trained Encoders Against Adversarial Examplescited by 21
  6. 2025ADVEDM:Fine-grained Adversarial Attack against VLM-based Embodied Agentscited by 19
  7. 2024NumbOD: A Spatial-Frequency Fusion Attack Against Object Detectorscited by 17
  8. 2024PB-UAP: Hybrid Universal Adversarial Attack For Image Segmentationcited by 14
  9. 2024Breaking Barriers in Physical-World Adversarial Examples: Improving Robustness and Transferability via Robust Featurecited by 13
  10. 2025Vanish into Thin Air: Cross-prompt Universal Adversarial Attacks for SAM2cited by 11
  11. 2025SegTrans: Transferable Adversarial Examples for Segmentation Modelscited by 8
  12. 2024VistaFormer: Scalable Vision Transformers for Satellite Image Time Series Segmentationcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026Transferable Physical-World Adversarial Patches Against Pedestrian Detection Modelscited by 2
  2. 2026Transferable Physical-World Adversarial Patches Against Object Detection in Autonomous Drivingcited by 1
  3. 2026VFACamou: View-Fused Adversarial Camouflage for Environment-Adaptive Physical Evasioncited by 0