Targeted Attack for Deep Hashing Based Retrieval

Bai, Jiawang; Chen, Bin; Li, Yiming; Wu, Dongxian; Guo, Weiwei; Xia, Shu-Tao; Yang, En-hui

doi:10.1007/978-3-030-58452-8_36

Cited by 66 publications

(61 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The method in [261] can match the predicted distance to a different target scene or directly fabricate the depth of specific instances in the scene. Targeted attacks on hashing based retrieval are proposed in [262], [119], whereas a universal perturbation for image retrieval systems is computed in [263]. An example of adversarial attack on Graph Matching can be found in [264].…”

Section: E Miscellaneous Attacksmentioning

confidence: 99%

Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey

2018

View full text Add to dashboard Cite

Deep learning is at the heart of the current rise of artificial intelligence. In the field of Computer Vision, it has become the workhorse for applications ranging from self-driving cars to surveillance and security. Whereas deep neural networks have demonstrated phenomenal success (often beyond human capabilities) in solving complex problems, recent studies show that they are vulnerable to adversarial attacks in the form of subtle perturbations to inputs that lead a model to predict incorrect outputs. For images, such perturbations are often too small to be perceptible, yet they completely fool the deep learning models. Adversarial attacks pose a serious threat to the success of deep learning in practice. This fact has recently lead to a large influx of contributions in this direction. This article presents the first comprehensive survey on adversarial attacks on deep learning in Computer Vision. We review the works that design adversarial attacks, analyze the existence of such attacks and propose defenses against them. To emphasize that adversarial attacks are possible in practical conditions, we separately review the contributions that evaluate adversarial attacks in the real-world scenarios. Finally, drawing on the reviewed literature, we provide a broader outlook of this research direction.

show abstract

Section: E Miscellaneous Attacksmentioning

confidence: 99%

Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey

2018

View full text Add to dashboard Cite

show abstract

“…For the attack on deep hashing based retrieval models, HAG [34] is the state-of-the-art untargeted method, it solved the vanishing gradient problem but it still used the idea of pushing the embedding of adversary query away from that of the original one. For the targeted attack, DHTA [3] improved [30] by calculating an anchor code and then move the adversarial query close to it.…”

Section: Adversarial Attack On Retrieval Systemsmentioning

confidence: 99%

“…In the right side figure, when (q) has the same () with ℎ(r ) (e.g., 1 for Case 1), our function wishes to change the () of (q) and assign a gradient on it. When (q) is around 0 (e.g., 3 ), its gradient is relatively larger than the others, and the gradient decreases to 0 fast when (q) • ℎ(r ) have different () (e.g., 2 ). However, the gradient of the inner product based function stay at the same value.…”

Section: Untargeted Adversarial Attackmentioning

confidence: 99%

“…We apply the same early termination condition (i.e., AP = 0) described in Section 4.4 on it, such that the algorithm can stop earlier and avoid unnecessarily large perturbations. • DHTA [3] is the state-of-the-art targeted deep hashing retrieval system attack. The core idea is to move the adversarial query close to a pre-computed anchor code h in the embedding space.…”

Section: Experiments 51 Datasets and Evaluation Protocolmentioning

confidence: 99%

“…Like other DNNs-based models (e.g., image classification), DNNsbased hashing retrieval models are also proved to be vulnerable to adversarial attacks [34]. This means the outputs of deep hashing retrieval models would be readily misled or affected by an adversarial imageq which added a small perturbation on the original query image q. Functionally, the adversarial attacks could be further classified into untargeted attack [26,34,38] and targeted attack [3,30], according to whether certain malicious content or labels are specified by the attackers.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Smart Adversarial Attack on Deep Hashing Based Image Retrieval

Chen

Sun

et al. 2021

Proceedings of the 2021 International Conference on Multimedia Retrieval

View full text Add to dashboard Cite

Deep hashing based retrieval models have been widely used in largescale image retrieval systems. Recently, there has been a surging interest in studying the adversarial attack problem in deep hashing based retrieval models. However, the effectiveness of existing adversarial attacks is limited by their poor perturbation management, unawareness of ranking weight, and only laser-focusing on the attack image. These shortages lead to high perturbation costs yet low AP reductions. To overcome these shortages, we propose a novel adversarial attack framework to improve the effectiveness of adversarial attacks. Our attack designs a dimension-wise surrogate Hamming distance function to help with wiser perturbation management. Further, in generating adversarial examples, instead of focusing on a single image, we propose to collectively incorporate relevant images combined with an AP-oriented (average precision) weight function. In addition, our attack can deal with both untargeted and targeted adversarial attacks in a flexible manner. Extensive experiments demonstrate that, with the same attack performance, our model significantly outperforms state-of-the-art models in perturbation cost on both untargeted and targeted attack tasks. CCS CONCEPTS• Computing methodologies → Computer vision problems; • Security and privacy → Software and application security.

show abstract

Sparse Adversarial Attack via Perturbation Factorization

Fan

Zhang

et al. 2020

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Targeted Attack for Deep Hashing Based Retrieval

Cited by 66 publications

References 42 publications

Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey

Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey

A Smart Adversarial Attack on Deep Hashing Based Image Retrieval

Sparse Adversarial Attack via Perturbation Factorization

Contact Info

Product

Resources

About