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

Although Deep Neural Networks (DNNs), such as the convolutional neuralnetworks (CNN) and Vision Transformers (ViTs), have been successfully appliedin the field of computer vision, they are demonstrated to be vulnerable towell-sought Adversarial Examples (AEs) that can easily fool the DNNs. Theresearch in AEs has been active, and many adversarial attacks and explanationshave been proposed since they were discovered in 2014. The mystery of the AE'sexistence is still an open question, and many studies suggest that DNN trainingalgorithms have blind spots. The salient objects usually do not overlap withboundaries; hence, the boundaries are not the DNN model's attention.Nevertheless, recent studies show that the boundaries can dominate the behaviorof the DNN models. Hence, this study aims to look at the AEs from a differentperspective and proposes an imperceptible adversarial attack that systemicallyattacks the input image boundary for finding the AEs. The experimental resultshave shown that the proposed boundary attacking method effectively attacks sixCNN models and the ViT using only 32% of the input image content (from theboundaries) with an average success rate (SR) of 95.2% and an average peaksignal-to-noise ratio of 41.37 dB. Correlation analyses are conducted,including the relation between the adversarial boundary's width and the SR andhow the adversarial boundary changes the DNN model's attention. This paper'sdiscoveries can potentially advance the understanding of AEs and provide adifferent perspective on how AEs can be constructed.

Imperceptible Adversarial Attack on Deep Neural Networks from Image Boundary