Dynamic Perturbation-Adaptive Adversarial Training on Medical Image Classification

arXiv:2403.06798v1 Announce Type: cross
Abstract: Remarkable successes were made in Medical Image Classification (MIC) recently, mainly due to wide applications of convolutional neural networks (CNNs). However, adversarial examples (AEs) exhibited imperceptible similarity with raw data, raising serious concerns on network robustness. Although adversarial training (AT), in responding to malevolent AEs, was recognized as an effective approach to improve robustness, it was challenging to overcome generalization decline of networks caused by the AT. In this paper, in order to reserve high …

abstract adversarial adversarial examples adversarial training applications arxiv classification cnns concerns convolutional neural networks cs.cv cs.lg data dynamic eess.iv examples however image medical network networks neural networks raw robustness training type