I'm involved in OCR, and would like to use a large dataset of printed characters (not handwritten). Is there such a dataset available? It would be nice to find one having different fonts and/or noisy images.

Similar questions and discussions