Clean Image May be Dangerous: Data Poisoning Attacks Against Deep Hashing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Shuai, Zhang, Jie, Qi, Yuang, Chen, Kejiang, Zhang, Tianwei, Zhang, Weiming, Yu, Nenghai
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912297322020864
author Li, Shuai
Zhang, Jie
Qi, Yuang
Chen, Kejiang
Zhang, Tianwei
Zhang, Weiming
Yu, Nenghai
author_facet Li, Shuai
Zhang, Jie
Qi, Yuang
Chen, Kejiang
Zhang, Tianwei
Zhang, Weiming
Yu, Nenghai
contents Large-scale image retrieval using deep hashing has become increasingly popular due to the exponential growth of image data and the remarkable feature extraction capabilities of deep neural networks (DNNs). However, deep hashing methods are vulnerable to malicious attacks, including adversarial and backdoor attacks. It is worth noting that these attacks typically involve altering the query images, which is not a practical concern in real-world scenarios. In this paper, we point out that even clean query images can be dangerous, inducing malicious target retrieval results, like undesired or illegal images. To the best of our knowledge, we are the first to study data \textbf{p}oisoning \textbf{a}ttacks against \textbf{d}eep \textbf{hash}ing \textbf{(\textit{PADHASH})}. Specifically, we first train a surrogate model to simulate the behavior of the target deep hashing model. Then, a strict gradient matching strategy is proposed to generate the poisoned images. Extensive experiments on different models, datasets, hash methods, and hash code lengths demonstrate the effectiveness and generality of our attack method.
format Preprint
id arxiv_https___arxiv_org_abs_2503_21236
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Clean Image May be Dangerous: Data Poisoning Attacks Against Deep Hashing
Li, Shuai
Zhang, Jie
Qi, Yuang
Chen, Kejiang
Zhang, Tianwei
Zhang, Weiming
Yu, Nenghai
Computer Vision and Pattern Recognition
Multimedia
Large-scale image retrieval using deep hashing has become increasingly popular due to the exponential growth of image data and the remarkable feature extraction capabilities of deep neural networks (DNNs). However, deep hashing methods are vulnerable to malicious attacks, including adversarial and backdoor attacks. It is worth noting that these attacks typically involve altering the query images, which is not a practical concern in real-world scenarios. In this paper, we point out that even clean query images can be dangerous, inducing malicious target retrieval results, like undesired or illegal images. To the best of our knowledge, we are the first to study data \textbf{p}oisoning \textbf{a}ttacks against \textbf{d}eep \textbf{hash}ing \textbf{(\textit{PADHASH})}. Specifically, we first train a surrogate model to simulate the behavior of the target deep hashing model. Then, a strict gradient matching strategy is proposed to generate the poisoned images. Extensive experiments on different models, datasets, hash methods, and hash code lengths demonstrate the effectiveness and generality of our attack method.
title Clean Image May be Dangerous: Data Poisoning Attacks Against Deep Hashing
topic Computer Vision and Pattern Recognition
Multimedia
url https://arxiv.org/abs/2503.21236