PBCAT: Patch-based composite adversarial training against physically realizable attacks on object detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Xiao, Zhu, Yiming, Huang, Yifan, Zhang, Wei, He, Yingzhe, Shi, Jie, Hu, Xiaolin
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908441541345280
author Li, Xiao
Zhu, Yiming
Huang, Yifan
Zhang, Wei
He, Yingzhe
Shi, Jie
Hu, Xiaolin
author_facet Li, Xiao
Zhu, Yiming
Huang, Yifan
Zhang, Wei
He, Yingzhe
Shi, Jie
Hu, Xiaolin
contents Object detection plays a crucial role in many security-sensitive applications. However, several recent studies have shown that object detectors can be easily fooled by physically realizable attacks, \eg, adversarial patches and recent adversarial textures, which pose realistic and urgent threats. Adversarial Training (AT) has been recognized as the most effective defense against adversarial attacks. While AT has been extensively studied in the $l_\infty$ attack settings on classification models, AT against physically realizable attacks on object detectors has received limited exploration. Early attempts are only performed to defend against adversarial patches, leaving AT against a wider range of physically realizable attacks under-explored. In this work, we consider defending against various physically realizable attacks with a unified AT method. We propose PBCAT, a novel Patch-Based Composite Adversarial Training strategy. PBCAT optimizes the model by incorporating the combination of small-area gradient-guided adversarial patches and imperceptible global adversarial perturbations covering the entire image. With these designs, PBCAT has the potential to defend against not only adversarial patches but also unseen physically realizable attacks such as adversarial textures. Extensive experiments in multiple settings demonstrated that PBCAT significantly improved robustness against various physically realizable attacks over state-of-the-art defense methods. Notably, it improved the detection accuracy by 29.7\% over previous defense methods under one recent adversarial texture attack.
format Preprint
id arxiv_https___arxiv_org_abs_2506_23581
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PBCAT: Patch-based composite adversarial training against physically realizable attacks on object detection
Li, Xiao
Zhu, Yiming
Huang, Yifan
Zhang, Wei
He, Yingzhe
Shi, Jie
Hu, Xiaolin
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Object detection plays a crucial role in many security-sensitive applications. However, several recent studies have shown that object detectors can be easily fooled by physically realizable attacks, \eg, adversarial patches and recent adversarial textures, which pose realistic and urgent threats. Adversarial Training (AT) has been recognized as the most effective defense against adversarial attacks. While AT has been extensively studied in the $l_\infty$ attack settings on classification models, AT against physically realizable attacks on object detectors has received limited exploration. Early attempts are only performed to defend against adversarial patches, leaving AT against a wider range of physically realizable attacks under-explored. In this work, we consider defending against various physically realizable attacks with a unified AT method. We propose PBCAT, a novel Patch-Based Composite Adversarial Training strategy. PBCAT optimizes the model by incorporating the combination of small-area gradient-guided adversarial patches and imperceptible global adversarial perturbations covering the entire image. With these designs, PBCAT has the potential to defend against not only adversarial patches but also unseen physically realizable attacks such as adversarial textures. Extensive experiments in multiple settings demonstrated that PBCAT significantly improved robustness against various physically realizable attacks over state-of-the-art defense methods. Notably, it improved the detection accuracy by 29.7\% over previous defense methods under one recent adversarial texture attack.
title PBCAT: Patch-based composite adversarial training against physically realizable attacks on object detection
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2506.23581