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/

Global: Zendy Tools annual

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

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

While machine learning applications are getting mainstream owing to ademonstrated efficiency in solving complex problems, they suffer from inherentvulnerability to adversarial attacks. Adversarial attacks consist of additivenoise to an input which can fool a detector. Recently, successful real-worldprintable adversarial patches were proven efficient against state-of-the-artneural networks. In the transition from digital noise based attacks toreal-world physical attacks, the myriad of factors affecting object detectionwill also affect adversarial patches. Among these factors, view angle is one ofthe most influential, yet under-explored. In this paper, we study the effect ofview angle on the effectiveness of an adversarial patch. To this aim, wepropose the first approach that considers a multi-view context by combiningexisting adversarial patches with a perspective geometric transformation inorder to simulate the effect of view angle changes. Our approach has beenevaluated on two datasets: the first dataset which contains most real worldconstraints of a multi-view context, and the second dataset which empiricallyisolates the effect of view angle. The experiments show that view anglesignificantly affects the performance of adversarial patches, where in somecases the patch loses most of its effectiveness. We believe that these resultsmotivate taking into account the effect of view angles in future adversarialattacks, and open up new opportunities for adversarial defenses.

Adversarial Attacks in a Multi-view Setting: An Empirical Study of the Adversarial Patches Inter-view Transferability