AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Frequency Domain Model Augmentation for Adversarial Attack

Yuyang Long, Qilong Zhang, Boheng Zeng, Lianli Gao, Xianglong Liu, Jian Zhang, Jingkuan Song

ECCV 2022 · Adversarial robustness and security › Black-box and transfer attacks · approach: Transfer attacks on face recognition · 239 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add spectrum transformation (frequency-domain augmentation) to MI-FGSMMI-FGSM · Normal models, crafted on Inc-v3avg success rate: 54.5 -> 63.9 (S2I-FGSM); 81.7 (FI-MI) -> 86.6Improved
add spectrum transformation to TI-DIMTI-DIM · Nine defenses, crafted on Inc-v3avg success rate: 39.0 -> 74.7Improved
add spectrum transformation to SI-TI-DIMS2I-SI-TI-DIM · Nine defenses, ensembleavg success rate: 89.8 -> 95.4Improved
apply transformation in frequency domain rather than spatial domain (ablation)S2I-FGSM · Normal models, Inc-v3success rate: +15.0%Improved
add Gaussian noise xi and tuning matrix M (ablation)spectrum transformation · Six normal modelsavg success rateImproved
apply DCT on whole image instead of small blocks (ablation)S2I-FGSM · Inc-v3success rateImproved
increase number of spectrum transformations N from 1 to 20 (ablation)S2I-FGSM · Inc-v3success rateImproved
increase tuning factor rho beyond 0.5 (ablation)S2I-FGSM · Inc-v3success rateHurt
increase noise std sigma beyond 16 (ablation)S2I-FGSM · Inc-v3success rateHurt
combine spectrum transformation with FI/VT-style MI-FGSMFI-MI-FGSM · Six normal modelsavg success rate: 81.7 -> 86.6Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy