Ocropy OCR – good results for Gothic/Fraktur Typeface

My favorite feature of ocropy is the ocropus-gtedit module. After a recognition you can transform the data into a HTML-File (see fig. 3).

HTML from ocropus-gtedit
Figure 3: HTML for correcting recognised text.

This offers a very user-friendly possibility to correct the text and generate ground truth data for training.

There is some literature about ocropy as well. Check out Springmann 2016, Springmann 2015, and DANVK 2015.


Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.