Work involving not only historic but also modern prints and handwriting can be considerably simplified if there is a machine-readable and searchable full text. This can be produced using OCR (Optical Character Recognition) and HTR (Handwritten Text Recognition).

One program that can be used for this is the transcription platform Transkribus. Good to very good results can often be achieved from printed materials in particular, but also some forms of handwriting such as the German cursive known as Kurrent.

Costs

Each account receives 100 free credits per month. This allows approx. 100 pages of manuscript or approx. 600 pages of printed text to be OCRd.

Bachelor, Master and PhD students and course organizers may apply to have costs covered by a grant.

Various subscription plans are also offered. Please contact us for more information, support with funding for projects, etc.

Selecting a text recognition package

When selecting a public text recognition package the following criteria are relevant: language, period and type of script.

What to do if you do not know what type of script it is: Here you can find some examples with suggestions for potentially suitable text recognition packages. Results vary depending on how similar the script in the individual document is to the material on which the text recognition package has been trained.

Further information on the subject of automatic text recognition and other software can be found on the homepage of the OCR competence center operated jointly by the University Libraries of Tübingen and Mannheim. The OCR-Recommender will generate a recommendation for the best OCR method for you or simply send us some sample pages.

If you have any questions about the application or if you are interested in integrating automatic text recognition methods into your studies, your scientific work or a project, please take part in our open OCR consultation hours or contact Dorothee Huff.