Adversarial Attacks on Image Classification Models: FGSM and Patch Attacks and their Impact
Abstract
This chapter introduces the concept of adversarial attacks on image classification models built on convolutional neural networks (CNN). CNNs are very popular deep-learning models which are used in image classification tasks. However, very powerful and pre-trained CNN models working very accurately on image datasets for image classification tasks may perform disastrously when the networks are under adversarial attacks. In this work, two very well-known adversarial attacks are discussed and their impact on the performance of image classifiers is analyzed. These two adversarial attacks are the fast gradient sign method (FGSM) and adversarial patch attack. These attacks are launched on three powerful pre-trained image classifier architectures, ResNet-34, GoogleNet, and DenseNet-161. The classification accuracy of the models in the absence and presence of the two attacks are computed on images from the publicly accessible ImageNet dataset. The results are analyzed to evaluate the impact of the attacks on the image classification task.
Cite
@article{arxiv.2307.02055,
title = {Adversarial Attacks on Image Classification Models: FGSM and Patch Attacks and their Impact},
author = {Jaydip Sen and Subhasis Dasgupta},
journal= {arXiv preprint arXiv:2307.02055},
year = {2023}
}
Comments
This is the preprint of the chapter titled "Adversarial Attacks on Image Classification Models: FGSM and Patch Attacks and their Impact" which will be published in the volume titled "Information Security and Privacy in the Digital World - Some Selected Cases", edited by Jaydip Sen. The book will be published by IntechOpen, London, UK, in 2023. This is not the final version of the chapter