Study on adversarial attack vulnerability of medical image analysis systems published in Medical Image Analysis
Cristina González Gonzalo (Radboudumc/UvA), Gerda Bortsova (Erasmus MC) and Suzanne Wetstein (TU Eindhoven) study previously unexplored factors affecting adversarial attack vulnerability of deep learning medical image analysis systems in three medical domains: ophthalmology, radiology, and pathology.