The understanding of anatomical shapes as well as their variations over time or among individuals is crucial in medical diagnostics and therapy. Riemannian shape statistics in combination with deep learning may lead to novel, data-driven approaches for assessment of anomalies and morphological scoring in medicine.