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 Plus annual

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

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Adversarial attacks have highlighted the vulnerability of classifiers basedon machine learning for Synthetic Aperture Radar (SAR) Automatic TargetRecognition (ATR) tasks. An adversarial attack perturbs SAR images of on-groundtargets such that the classifiers are misled into making incorrect predictions.However, many existing attacking techniques rely on arbitrary manipulation ofSAR images while overlooking the feasibility of executing the attacks onreal-world SAR imagery. Instead, adversarial attacks should be able to beimplemented by physical actions, for example, placing additional false objectsas scatterers around the on-ground target to perturb the SAR image and fool theSAR ATR.  In this paper, we propose the On-Target Scatterer Attack (OTSA), ascatterer-based physical adversarial attack. To ensure the feasibility of itsphysical execution, we enforce a constraint on the positioning of thescatterers. Specifically, we restrict the scatterers to be placed only on thetarget instead of in the shadow regions or the background. To achieve this, weintroduce a positioning score based on Gaussian kernels and formulate anoptimization problem for our OTSA attack. Using a gradient ascent method tosolve the optimization problem, the OTSA can generate a vector of parametersdescribing the positions, shapes, sizes and amplitudes of the scatterers toguide the physical execution of the attack that will mislead SAR imageclassifiers. The experimental results show that our attack obtainssignificantly higher success rates under the positioning constraint comparedwith the existing method.

Realistic Scatterer Based Adversarial Attacks on SAR Image Classifiers