Unsupervised Multi-Criteria Adversarial Detection in Deep Image Retrieval
This addresses security challenges in image retrieval systems, offering a domain-specific defense against adversarial attacks, though it is incremental as it adapts detection strategies to this context.
The paper tackles the vulnerability of deep image retrieval systems to adversarial attacks by proposing an unsupervised scheme that detects adversarial behaviors using three criteria in hamming space, resulting in 2-23% improvements in detection rates across four datasets.
The vulnerability in the algorithm supply chain of deep learning has imposed new challenges to image retrieval systems in the downstream. Among a variety of techniques, deep hashing is gaining popularity. As it inherits the algorithmic backend from deep learning, a handful of attacks are recently proposed to disrupt normal image retrieval. Unfortunately, the defense strategies in softmax classification are not readily available to be applied in the image retrieval domain. In this paper, we propose an efficient and unsupervised scheme to identify unique adversarial behaviors in the hamming space. In particular, we design three criteria from the perspectives of hamming distance, quantization loss and denoising to defend against both untargeted and targeted attacks, which collectively limit the adversarial space. The extensive experiments on four datasets demonstrate 2-23% improvements of detection rates with minimum computational overhead for real-time image queries.