arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2409.01282cs.CVcs.CRcs.LG
收藏

基于单索引矢量量化的图像分类对抗攻击

One-Index Vector Quantization Based Adversarial Attack on Image Classification

Henan Normal University(河南师范大学);Durham University(杜伦大学)

Haiju Fan, Xiaona Qin, Shuang Chen, Hubert P. H. Shum, Ming Li

AI总结:

本文提出一种基于矢量量化域的单索引对抗攻击方法,通过差分进化仅修改一个压缩索引即可使Resnet、NIN和VGG16等模型对CIFAR-10及Fashion MNIST图像产生高置信度错误分类。

摘要:

为了改善存储和传输,图像通常会被压缩。矢量量化(VQ)是一种流行的压缩方法,因为它具有高压缩比,优于其他压缩技术。尽管如此,现有的图像分类对抗攻击方法大多在像素域进行,只有少数例外在压缩域进行,这使得它们在现实场景中适用性较低。本文提出了一种新颖的VQ域单索引攻击方法,通过差分进化算法生成对抗图像,成功导致受害模型中的图像错误分类。该单索引攻击方法修改压缩数据流中的单个索引,使解压后的图像被错误分类。它只需要修改单个VQ索引即可实现攻击,从而限制了被扰动索引的数量。所提方法属于半黑盒攻击,更符合实际攻击场景。我们将该方法应用于攻击三种流行的图像分类模型,即Resnet、NIN和VGG16。平均而言,CIFAR-10和Fashion MNIST中分别有55.9%和77.4%的图像被成功攻击,具有较高的错误分类置信度和较低的图像扰动水平。

英文摘要:

To improve storage and transmission, images are generally compressed. Vector quantization (VQ) is a popular compression method as it has a high compression ratio that suppresses other compression techniques. Despite this, existing adversarial attack methods on image classification are mostly performed in the pixel domain with few exceptions in the compressed domain, making them less applicable in real-world scenarios. In this paper, we propose a novel one-index attack method in the VQ domain to generate adversarial images by a differential evolution algorithm, successfully resulting in image misclassification in victim models. The one-index attack method modifies a single index in the compressed data stream so that the decompressed image is misclassified. It only needs to modify a single VQ index to realize an attack, which limits the number of perturbed indexes. The proposed method belongs to a semi-black-box attack, which is more in line with the actual attack scenario. We apply our method to attack three popular image classification models, i.e., Resnet, NIN, and VGG16. On average, 55.9% and 77.4% of the images in CIFAR-10 and Fashion MNIST, respectively, are successfully attacked, with a high level of misclassification confidence and a low level of image perturbation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑