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

Cited by in corpus (11)

• Adversarial attacks on Copyright Detection Systems
• Stabilized Medical Image Attacks
• 3DB: A Framework for Debugging Computer Vision Models
• 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
• Efficient Adversarial Attacks for Visual Object Tracking
• Temporally-Transferable Perturbations: Efficient, One-Shot Adversarial Attacks for Online Visual Object Trackers
• Provably robust deep generative models
• Intelligent image synthesis to attack a segmentation CNN using adversarial learning
• Landmark Breaker: Obstructing DeepFake By Disturbing Landmark Extraction
• Correlation between image quality metrics of magnetic resonance images and the neural network segmentation accuracy