Understanding adversarial attacks on deep learning based medical image analysis systems

Deep neural networks (DNNs) have become popular for medical image analysis tasks like cancer diagnosis and lesion detection. However, a recent study demonstrates that deep learning systems can be compromised by carefully-engineered adversarial examples/attacks with small imperceptible perturbations. This raises safety concerns about the deployment of these in clinical settings. In this paper, we provide deeper understanding examples context images. We find DNN models more vulnerable to attacks compared natural images, according two different viewpoints. Surprisingly, also easily detected, i.e., simple detectors achieve over 98% detection AUC against state-of-the-art attacks, due fundamental feature differences normal examples. believe findings may useful basis approach design explainable secure systems.