Evaluating OCR and Non-OCR Text Representations for Learning Document Classifiers.

Markus Junker, Rainer Hoch

Browse the full ICDAR paper archive.