Radiol Artif Intell 2021 May 17;3(3):e190169. Epub 2021 Feb 17.
Department of Computer Aided Medical Procedures, Technical University of Munich, Boltzmannstr 3, 85748 Garching near Munich, Germany (C.B., N.N., S.A.); Department of Diagnostic and Interventional Neuroradiology (B.W., C.Z.) and Department of Neurology (M.M.), Klinikum Rechts der Isar, Technical University of Munich, Munich, Germany; and Whiting School of Engineering, Johns Hopkins University, Baltimore, Md (N.N.).
Purpose: To develop an unsupervised deep learning model on MR images of normal brain anatomy to automatically detect deviations indicative of pathologic states on abnormal MR images.
Materials And Methods: In this retrospective study, spatial autoencoders with skip-connections (which can learn to compress and reconstruct data) were leveraged to learn the normal variability of the brain from MR scans of healthy individuals. A total of 100 normal, in-house MR scans were used for training. Read More