Frequency constraint-based adversarial attack on deep neural networks for medical image classification

Fang Chen, Jian Wang, Han Liu, Wentao Kong, Zhe Zhao, Longfei Ma, Hongen Liao, Daoqiang Zhang

Abstract: Highlights•A novel frequency constraint-based adversarial attack is developed for various medical image classification tasks.•The low-frequency constraints restrict the perturbations to the high-frequency components and promotes imperceptibility.•Experiments on four public medical image datasets demonstrate the superior performance of our model over SOTA methods.