Autoencoder for processing simple geometric attributes

The dataset used in the paper is a collection of grey-level images of centred disks with varying radii.

BibTex: