Question
How to recognize Gothic (Fraktur) text using FineReader PDF?
Answer
The following languages are supported for the recognition of Gothic (Fraktur) text in FineReader PDF:
- Old English (dictionary support)
- Old French (dictionary support)
- Old German (dictionary support)
- Old Italian (dictionary support)
- Old Spanish (dictionary support)
- Latvian Gothic
To recognize Gothic text in FineReader PDF, please follow the steps below:
In FineReader PDF 16:
- Run FineReader PDF.
- Go to File menu > Options... > Languages.
- Select the Specify OCR languages manually option.
- Select the required language in the OCR Languages dialog box > click OK.
E.g. for German Fraktur select Old German.
In FineReader PDF 15:
- Run FineReader PDF 15.
- Go to Tools menu > Options... > Languages.
- Select the Specify OCR languages manually option.
- Select the required language in the OCR Languages dialog box > click OK.
The recognition quality of text printed in Gothic is better if the only Gothic language is selected. In cases, where 'Gothic' and 'standard' languages are mixed, the recognition quality might decrease for texts written in Gothic. To achieve the highest recognition quality, it is recommended to select only the Gothic language for recognition of Gothic text within a document.
Comments
4 comments
Michael Preuscchoff
Ich will einen Frakturschrifttext scannen, doch wenn ich finereader PDF 15 öffne, steht da "Müheloses arbeiten mit PDFS" usw., doch auf frakturerkennung weist darauf NICHTS hin. Ich finde, hier ist NICHTS selbsterklärend!
I want to scan a text in gothic letters, but there is not any indication how to manage this!
Victoria Dvornikova
Hello Michael Preuscchoff,
I have created a support ticket based on your comment. Please await a reply from our Customer Support agent.
steuber
Hello, I need to OCR a document with Fraktur font (a Schwabacher). The default setting Alt-German brings only very poor results. I then tried to create my own pattern. This works better, but there is a problem with the lowercase letter "d" that Finereader consistently interprets as "b", even after I have entered at least 30 times a "d" of the scan also as "d" in the pattern recognition. Could it be that there is an error in the programming? b and d differ quite clearly in the Schwabacher.
It would be good to get a really useful tutorial for the recognition of Fraktur and the creation of own patterns. For example, I don't know if it makes sense at all to train the same letters over and over again or if it is more useful to train only one or a few patterns as letters.
I also can't apply a once trained pattern to other parts of the text in the document, that produces bad results again despite the same font.
It would be really useful to get a small tutorial using a real historical example document. Would that be possible?
Victoria Dvornikova
Hello steuber,
I have created a support ticket based on your comment. Please await a reply from our Customer support agent.
Please sign in to leave a comment.