RUS  ENG
Full version
JOURNALS // Zapiski Nauchnykh Seminarov POMI // Archive

Zap. Nauchn. Sem. POMI, 2023 Volume 530, Pages 128–140 (Mi znsl7437)

This article is cited in 1 paper

Realistic adversarial attacks on object detectors using generative models

D. Shelepneva, K. Arkhipenko

Ivannikov Institute for System Programming of the RAS

Abstract: An important limitation of existing adversarial attacks on real-world object detectors lies in their threat model: adversarial patch-based methods often produce suspicious images while image generation approaches do not restrict the attacker's capabilities of modifying the original scene. We design a threat model where the attacker modifies individual image segments and is required to produce realistic images. We also develop and evaluate a white-box attack that utilizes generative adversarial nets and diffusion models as a generator of malicious images. Our attack is able to produce high-fidelity images as measured by the Fréchet inception distance (FID) and reduces the mAP of Faster R-CNN model by > 0.2 on Cityscapes and COCO-Stuff datasets. A PyTorch implementation of our attack is available at https://github.com/DariaShel/gan-attack.

Key words and phrases: adversarial examples, object detectors, generative adversarial networks, diffusion models.

UDC: 004.852

Received: 06.09.2023

Language: English


 English version:
Journal of Mathematical Sciences (New York), 2024, 285:2, 245–254


© Steklov Math. Inst. of RAS, 2025