mlsanityData to 7 Oct 2026

ConSeg: Contextual Backdoor Attack Against Semantic Segmentation

arXiv 2507.19905 ↗First posted 26 Jul 2025cs.CR cs.CVcited by 0

On the map: LLM safety and security → Backdoor and data poisoning attacks

Abstract (arXiv)

Despite significant advancements in computer vision, semantic segmentation models may be susceptible to backdoor attacks. These attacks, involving hidden triggers, aim to cause the models to misclassify instances of the victim class as the target class when triggers are present, posing serious threats to the reliability of these models. To further explore the field of backdoor attacks against semantic segmentation, in this paper, we propose a simple yet effective backdoor attack called Contextual Segmentation Backdoor Attack (ConSeg). ConSeg leverages the contextual information inherent in semantic segmentation models to enhance backdoor performance. Our method is motivated by an intriguing observation, i.e., when the target class is set as the `co-occurring' class of the victim class, the victim class can be more easily `mis-segmented'. Building upon this insight, ConSeg mimics the contextual information of the target class and rebuilds it in the victim region to establish the contextual relationship between the target class and the victim class, making the attack easier. Our experiments reveal that ConSeg achieves improvements in Attack Success Rate (ASR) with increases of 15.55%, compared to existing methods, while exhibiting resilience against state-of-the-art backdoor defenses.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Backdoor and data poisoning attacks, one of 12 directions in LLM safety and security. Builds on 44 of its 60 references that are in the map. First posted on arXiv 26 Jul 2025.

Builds on

44 of 60 references.

Closest ancestors · 1

Same direction, newest first

  1. 2021Rethinking the Backdoor Attacks' Triggers: A Frequency Perspective7 Apr

Other references · 43

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  2. 2023GPT-4 Technical Reportcited by 18,132
  3. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
  4. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,845
  5. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
  6. 2016DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFscited by 3,239
  7. 2016Pyramid Scene Parsing Networkcited by 2,472
  8. 2018BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learningcited by 1,282
  9. 2017Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoningcited by 1,012
  10. 2018Dual Attention Network for Scene Segmentationcited by 1,001
  11. 2018Fine-Pruning: Defending Against Backdooring Attacks on Deep Neural Networkscited by 550
  12. 2019STRIP: A Defence Against Trojan Attacks on Deep Neural Networkscited by 437
31 more
  1. 2019Hidden Trigger Backdoor Attackscited by 359
  2. 2020Invisible Backdoor Attack with Sample-Specific Triggerscited by 338
  3. 2019Deep Multi-modal Object Detection and Semantic Segmentation for Autonomous Driving: Datasets, Methods, and Challengescited by 313
  4. 2020Attack of the Tails: Yes, You Really Can Backdoor Federated Learningcited by 307
  5. 2017Adversarial Examples for Semantic Segmentation and Object Detectioncited by 275
  6. 2020PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentationcited by 274
  7. 2021Adversarial Neuron Pruning Purifies Backdoored Deep Modelscited by 230
  8. 2017Neural Trojanscited by 151
  9. 2022Backdoor Defense via Decoupling the Training Processcited by 149
  10. 2020Dynamic Backdoor Attacks Against Machine Learning Modelscited by 147
  11. 2020Bridging Mode Connectivity in Loss Landscapes and Adversarial Robustnesscited by 130
  12. 2017On the Robustness of Semantic Segmentation Models to Adversarial Attackscited by 126
  13. 2019Deep Semantic Segmentation of Natural and Medical Images: A Reviewcited by 124
  14. 2020RAB: Provable Robustness Against Backdoor Attackscited by 115
  15. 2022Data-free Backdoor Removal based on Channel Lipschitznesscited by 77
  16. 2022SegPGD: An Effective and Efficient Adversarial Attack for Evaluating and Boosting Segmentation Robustnesscited by 74
  17. 2021Handcrafted Backdoors in Deep Neural Networkscited by 57
  18. 2022BadDet: Backdoor Attacks on Object Detectioncited by 45
  19. 2023Detecting Backdoors During the Inference Stage Based on Corruption Robustness Consistencycited by 45
  20. 2018A Curriculum Domain Adaptation Approach to the Semantic Segmentation of Urban Scenescited by 43
  21. 2022On the Real-World Adversarial Robustness of Real-Time Semantic Segmentation Models for Autonomous Drivingcited by 30
  22. 2022Imperceptible Backdoor Attack: From Input Space to Feature Representationcited by 27
  23. 2024Towards Robust Physical-world Backdoor Attacks on Lane Detectioncited by 25
  24. 2022Dangerous Cloaking: Natural Trigger based Backdoor Attacks on Object Detectors in the Physical Worldcited by 20
  25. 2021Hidden Backdoor Attack against Semantic Segmentation Modelscited by 19
  26. 2019AdvSPADE: Realistic Unrestricted Attacks for Semantic Segmentationcited by 10
  27. 2023Watch Out! Simple Horizontal Class Backdoor Can Trivially Evade Defensecited by 7
  28. 2023SoK: A Systematic Evaluation of Backdoor Trigger Characteristics in Image Classificationcited by 6
  29. 2023Influencer Backdoor Attack on Semantic Segmentationcited by 5
  30. 2023Robust Backdoor Attacks on Object Detection in Real Worldcited by 5
  31. 2024BadFusion: 2D-Oriented Backdoor Attacks against 3D Object Detectioncited by 2

Built on by

Not cited yet.