Universal Physical Camouflage Attacks on Object Detectors

TL;DR

This paper introduces Universal Physical Camouflage Attacks (UPC) that craft adversarial patterns to fool object detectors across all instances of a category, effective in both virtual and real-world environments.

Contribution

We propose a novel universal camouflage attack method that generalizes to non-rigid objects and introduces a standardized virtual testing environment for evaluation.

Findings

UPC outperforms existing physical attacks in virtual environments.

UPC effectively fools object detectors in real-world tests.

The attack maintains natural appearance to human observers.

Abstract

In this paper, we study physical adversarial attacks on object detectors in the wild. Previous works mostly craft instance-dependent perturbations only for rigid or planar objects. To this end, we propose to learn an adversarial pattern to effectively attack all instances belonging to the same object category, referred to as Universal Physical Camouflage Attack (UPC). Concretely, UPC crafts camouflage by jointly fooling the region proposal network, as well as misleading the classifier and the regressor to output errors. In order to make UPC effective for non-rigid or non-planar objects, we introduce a set of transformations for mimicking deformable properties. We additionally impose optimization constraint to make generated patterns look natural to human observers. To fairly evaluate the effectiveness of different physical-world attacks, we present the first standardized virtual database, AttackScenes, which simulates the real 3D world in a controllable and reproducible environment. Extensive experiments suggest the superiority of our proposed UPC compared with existing physical adversarial attackers not only in virtual environments (AttackScenes), but also in real-world physical environments. Code and dataset are available at https://mesunhlf.github.io/index_physical.html.