A Perturbation-Constrained Adversarial Attack for Evaluating the Robustness of Optical Flow

TL;DR

This paper introduces PCFA, a novel adversarial attack designed to evaluate the robustness of optical flow methods by generating worst-case perturbations, revealing their vulnerabilities beyond accuracy metrics.

Contribution

The paper presents PCFA, a global, perturbation-constrained attack that effectively assesses optical flow robustness, and provides the first combined ranking considering accuracy and adversarial resilience.

Findings

PCFA finds stronger adversarial samples than previous methods.

State-of-the-art optical flow methods are vulnerable to PCFA attacks.

The joint ranking highlights the robustness gaps in current methods.

Abstract

Recent optical flow methods are almost exclusively judged in terms of accuracy, while their robustness is often neglected. Although adversarial attacks offer a useful tool to perform such an analysis, current attacks on optical flow methods focus on real-world attacking scenarios rather than a worst case robustness assessment. Hence, in this work, we propose a novel adversarial attack - the Perturbation-Constrained Flow Attack (PCFA) - that emphasizes destructivity over applicability as a real-world attack. PCFA is a global attack that optimizes adversarial perturbations to shift the predicted flow towards a specified target flow, while keeping the L2 norm of the perturbation below a chosen bound. Our experiments demonstrate PCFA's applicability in white- and black-box settings, and show it finds stronger adversarial samples than previous attacks. Based on these strong samples, we provide the first joint ranking of optical flow methods considering both prediction quality and adversarial robustness, which reveals state-of-the-art methods to be particularly vulnerable. Code is available at https://github.com/cv-stuttgart/PCFA.