A Convolutional Autoencoder Approach to Learn Volumetric Shape Representations for Brain Structures | Read Paper on Bytez