On the Robustness of Semantic Segmentation Models to Adversarial Attacks

TL;DR

This paper rigorously evaluates the robustness of modern semantic segmentation models against adversarial attacks across large datasets, revealing insights into architecture vulnerabilities and potential defense mechanisms.

Contribution

It provides the first comprehensive analysis of adversarial robustness in semantic segmentation, comparing architectures and exploring inherent defense strategies.

Findings

Many classification observations do not transfer to segmentation.

Multiscale processing and input transformations can serve as defenses.

Certain models demonstrate higher robustness, guiding safer model selection.

Abstract

Deep Neural Networks (DNNs) have demonstrated exceptional performance on most recognition tasks such as image classification and segmentation. However, they have also been shown to be vulnerable to adversarial examples. This phenomenon has recently attracted a lot of attention but it has not been extensively studied on multiple, large-scale datasets and structured prediction tasks such as semantic segmentation which often require more specialised networks with additional components such as CRFs, dilated convolutions, skip-connections and multiscale processing. In this paper, we present what to our knowledge is the first rigorous evaluation of adversarial attacks on modern semantic segmentation models, using two large-scale datasets. We analyse the effect of different network architectures, model capacity and multiscale processing, and show that many observations made on the task of classification do not always transfer to this more complex task. Furthermore, we show how mean-field inference in deep structured models, multiscale processing (and more generally, input transformations) naturally implement recently proposed adversarial defenses. Our observations will aid future efforts in understanding and defending against adversarial examples. Moreover, in the shorter term, we show how to effectively benchmark robustness and show which segmentation models should currently be preferred in safety-critical applications due to their inherent robustness.