We plan to continue to identify and test deep-learning models for quantitative medical image analysis, in particular anatomical segmentation (for example FastSurfer, Henschel et al. 2020) and brain lesion segmentation (for example DeepMedic, Kamnitsas et al. (2017)). Alvis will be used for image preprocessing, training, validating, and finding optimal model parameters, bringing the models closer to clinical application.