English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Recent optical flow methods are almost exclusively judged in terms ofaccuracy, while their robustness is often neglected. Although adversarialattacks offer a useful tool to perform such an analysis, current attacks onoptical flow methods focus on real-world attacking scenarios rather than aworst case robustness assessment. Hence, in this work, we propose a noveladversarial attack - the Perturbation-Constrained Flow Attack (PCFA) - thatemphasizes destructivity over applicability as a real-world attack. PCFA is aglobal attack that optimizes adversarial perturbations to shift the predictedflow towards a specified target flow, while keeping the L2 norm of theperturbation below a chosen bound. Our experiments demonstrate PCFA'sapplicability in white- and black-box settings, and show it finds strongeradversarial samples than previous attacks. Based on these strong samples, weprovide the first joint ranking of optical flow methods considering bothprediction quality and adversarial robustness, which reveals state-of-the-artmethods to be particularly vulnerable. Code is available athttps://github.com/cv-stuttgart/PCFA.

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