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 studies have shown convolution neural networks (CNNs) for imagerecognition are vulnerable to evasion attacks with carefully manipulatedadversarial examples. Previous work primarily focused on how to generateadversarial examples closed to source images, by introducing pixel-levelperturbations into the whole or specific part of images. In this paper, wepropose an evasion attack on CNN classifiers in the context of License PlateRecognition (LPR), which adds predetermined perturbations to specific regionsof license plate images, simulating some sort of naturally formed spots (suchas sludge, etc.). Therefore, the problem is modeled as an optimization processsearching for optimal perturbation positions, which is different from previouswork that consider pixel values as decision variables. Notice that this is acomplex nonlinear optimization problem, and we use a genetic-algorithm basedapproach to obtain optimal perturbation positions. In experiments, we use theproposed algorithm to generate various adversarial examples in the form ofrectangle, circle, ellipse and spots cluster. Experimental results show thatthese adversarial examples are almost ignored by human eyes, but can foolHyperLPR with high attack success rate over 93%. Therefore, we believe thatthis kind of spot evasion attacks would pose a great threat to current LPRsystems, and needs to be investigated further by the security community.

Spot Evasion Attacks: Adversarial Examples for License Plate Recognition Systems with Convolutional Neural Networks