autodoc

Can an Autoencoder trained on images of documents generate useful features for classification of a very small dataset? We believe it can. This project checks that...