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

Compared with transferable untargeted attacks, transferable targetedadversarial attacks could specify the misclassification categories ofadversarial samples, posing a greater threat to security-critical tasks. In themeanwhile, 3D adversarial samples, due to their potential of multi-viewrobustness, can more comprehensively identify weaknesses in existing deeplearning systems, possessing great application value. However, the field oftransferable targeted 3D adversarial attacks remains vacant. The goal of thiswork is to develop a more effective technique that could generate transferabletargeted 3D adversarial examples, filling the gap in this field. To achievethis goal, we design a novel framework named TT3D that could rapidlyreconstruct from few multi-view images into Transferable Targeted 3D texturedmeshes. While existing mesh-based texture optimization methods computegradients in the high-dimensional mesh space and easily fall into local optima,leading to unsatisfactory transferability and distinct distortions, TT3Dinnovatively performs dual optimization towards both feature grid andMulti-layer Perceptron (MLP) parameters in the grid-based NeRF space, whichsignificantly enhances black-box transferability while enjoying naturalness.Experimental results show that TT3D not only exhibits superior cross-modeltransferability but also maintains considerable adaptability across differentrenders and vision tasks. More importantly, we produce 3D adversarial exampleswith 3D printing techniques in the real world and verify their robustperformance under various scenarios.

Towards Transferable Targeted 3D Adversarial Attack in the Physical World