Backdoor Attack on Hash-based Image Retrieval via Clean-label Data Poisoning
Kuofeng Gao (Tsinghua University),* Jiawang Bai (Tsinghua University), Bin Chen (Harbin Institute of Technology, Shenzhen), Dongxian Wu (the University of Tokyo), Shu-Tao Xia (Tsinghua University)
The 34th British Machine Vision Conference
Abstract
A backdoored deep hashing model is expected to behave normally on original query images and return the images with the target label when a specific trigger pattern presents. To this end, we propose the confusing perturbations-induced backdoor attack (CIBA). It injects a small number of poisoned images with the correct label into the training data, which makes the attack hard to be detected. To craft the poisoned images, we first propose the confusing perturbations to disturb the hashing code learning. As such, the hashing model can learn more about the trigger. The confusing perturbations are imperceptible and generated by optimizing the intra-class dispersion and inter-class shift in the Hamming space. We then employ the targeted adversarial patch as the backdoor trigger to improve the attack performance. We have conducted extensive experiments to verify the effectiveness of our proposed CIBA.Video
Citation
@inproceedings{Gao_2023_BMVC,
author = {Kuofeng Gao and Jiawang Bai and Bin Chen and Dongxian Wu and Shu-Tao Xia},
title = {Backdoor Attack on Hash-based Image Retrieval via Clean-label Data Poisoning},
booktitle = {34th British Machine Vision Conference 2023, {BMVC} 2023, Aberdeen, UK, November 20-24, 2023},
publisher = {BMVA},
year = {2023},
url = {https://papers.bmvc2023.org/0172.pdf}
}
Copyright © 2023
The British Machine Vision Conference is organised by