AI Paper Atlas

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

A Large-Scale Multiple-Objective Method for Black-Box Attack against Object Detection

Siyuan Liang, Longkang Li, Yanbo Fan, Xiaojun Jia, Jingzhi Li, Baoyuan Wu, Xiaochun Cao

ECCV 2022 · Adversarial robustness and security › Black-box and transfer attacks · approach: Query-efficient black-box attacks · 24 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
replace the attack objective with TP+FP (minimize true positives and maximize false positives) (ablation)PRFA · GFLmAP after attack (lower is better): 0.31 -> 0.27Improved
replace the attack objective with TP+FP (ablation)PRFA · YOLOXmAP after attack (lower is better): 0.31 -> 0.28Improved
initialize with skip-based gradient-prior perturbation (ablation)PRFA · GFLmAP after attack (lower is better): 0.31 -> 0.21Improved
initialize with skip-based gradient-prior perturbation (ablation)PRFA · YOLOXmAP after attack (lower is better): 0.31 -> 0.31No clear effect
replace PRFA search with a genetic algorithm over the whole image (ablation)PRFA with skip-based init · GFLmAP after attack (lower is better): 0.21 -> 0.25Hurt
replace PRFA search with a genetic algorithm over the whole image (ablation)PRFA with skip-based init · YOLOXmAP after attack (lower is better): 0.31 -> 0.42Hurt
add random subset selection (ablation)GA with skip-based init · GFLmAP after attack (lower is better): 0.25 -> 0.20Improved
add random subset selection (ablation)GA with skip-based init · YOLOXmAP after attack (lower is better): 0.42 -> 0.29Improved
add divide-and-conquer (ablation)GARS with skip-based init · GFLmAP after attack (lower is better): 0.20 -> 0.18Improved
add divide-and-conquer (ablation)GARS with skip-based init · YOLOXmAP after attack (lower is better): 0.29 -> 0.31Hurt
add chain-based perturbations to form a mixed initial population (ablation)GARSDC with skip-based init · GFLaverage queries (lower is better): 3037 -> 1838Improved
add chain-based perturbations to form a mixed initial population (ablation)GARSDC with skip-based init · YOLOXmAP after attack (lower is better): 0.31 -> 0.23Improved

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