Adversarial Examples for Semantic Image Segmentation

Abstract

Machine learning methods in general and Deep Neural Networks in particular have shown to be vulnerable to adversarial perturbations. So far this phenomenon has mainly been studied in the context of whole-image classification. In this contribution, we analyse how adversarial perturbations can affect the task of semantic segmentation. We show how existing adversarial attackers can be transferred to this task and that it is possible to create imperceptible adversarial perturbations that lead a deep network to misclassify almost all pixels of a chosen class while leaving network prediction nearly unchanged outside this class.

ICLR 2017 workshop submission

References in corpus (1)

• The Cityscapes Dataset for Semantic Urban Scene Understanding

Cited by in corpus (30)

• Mitigating Adversarial Effects Through Randomization
• DARTS: Deceiving Autonomous Cars with Toxic Signs
• Adversarial Examples for Semantic Segmentation and Object Detection
• Secure and Trustworthy Artificial Intelligence-Extended Reality (AI-XR) for Metaverses
• Advbox: a toolbox to generate adversarial examples that fool neural networks
• When and How to Fool Explainable Models (and Humans) with Adversarial Examples
• Defense against Universal Adversarial Perturbations
• Adversarial attacks on Copyright Detection Systems
• PAC-learning in the presence of evasion adversaries
• Stabilized Medical Image Attacks
• 3DB: A Framework for Debugging Computer Vision Models
• Metamorphic Relation Based Adversarial Attacks on Differentiable Neural Computer
• SegDiff: Image Segmentation with Diffusion Probabilistic Models
• Segmentations-Leak: Membership Inference Attacks and Defenses in Semantic Image Segmentation
• Fooling Vision and Language Models Despite Localization and Attention Mechanism
• Universal, transferable and targeted adversarial attacks
• Better the Devil you Know: An Analysis of Evasion Attacks using Out-of-Distribution Adversarial Examples
• Attacking and Defending Machine Learning Applications of Public Cloud
• Built-in Vulnerabilities to Imperceptible Adversarial Perturbations
• Efficient Adversarial Attacks for Visual Object Tracking
• Improving Model Robustness with Latent Distribution Locally and Globally
• Modularity Matters: Learning Invariant Relational Reasoning Tasks
• Temporally-Transferable Perturbations: Efficient, One-Shot Adversarial Attacks for Online Visual Object Trackers
• SAGE: Stealthy Attack GEneration for Cyber-Physical Systems
• Chameleon: A Semi-AutoML framework targeting quick and scalable development and deployment of production-ready ML systems for SMEs
• Provably robust deep generative models
• Correlation between image quality metrics of magnetic resonance images and the neural network segmentation accuracy
• Intelligent image synthesis to attack a segmentation CNN using adversarial learning
• A Le Cam Type Bound for Adversarial Learning and Applications
• Landmark Breaker: Obstructing DeepFake By Disturbing Landmark Extraction