Robust Training with Data Augmentation for Medical Imaging Classification

TL;DR

This paper introduces RTDA, a robust training algorithm with data augmentation, that enhances the reliability of medical imaging classifiers against adversarial attacks and distribution shifts, outperforming existing methods.

Contribution

The study presents RTDA, a novel training approach combining data augmentation techniques to improve robustness and generalization in medical image classification.

Findings

RTDA outperforms baseline methods in robustness against adversarial attacks.

RTDA maintains high accuracy on clean data.

RTDA improves generalization under distribution shifts.

Abstract

Deep neural networks are increasingly being used to detect and diagnose medical conditions using medical imaging. Despite their utility, these models are highly vulnerable to adversarial attacks and distribution shifts, which can affect diagnostic reliability and undermine trust among healthcare professionals. In this study, we propose a robust training algorithm with data augmentation (RTDA) to mitigate these vulnerabilities in medical image classification. We benchmark classifier robustness against adversarial perturbations and natural variations of RTDA and six competing baseline techniques, including adversarial training and data augmentation approaches in isolation and combination, using experimental data sets with three different imaging technologies (mammograms, X-rays, and ultrasound). We demonstrate that RTDA achieves superior robustness against adversarial attacks and improved generalization performance in the presence of distribution shift in each image classification task while maintaining high clean accuracy.