Form Recognizer cyrillic text in scanned document

Vladimir Mushkin 6

Hello.
I tried to use Form Recognizer from MS Azure - it is very good for me, except one thing, if using scanned PDF Cyrillic language is recognized as a Latin character sets ad it makes the output unusable. I was thinking about using this Cloud Service in my company but really need support with Cyrillic Characters

1 answer

Ramr-msft 17,621 Reputation points

2021-05-05T15:50:48.49+00:00

@Vladimir Mushkin Thanks for the question. Can you please add more details about the input and the recognized output that you are getting.
Please follow document for list of Language supported in Form Recognizer.

Languages-Locales: Serbian: One of: SR-BA (Cyrillic or Latin), SR-ME (Cyrillic or Latin), SR-RS (Cyrillic or Latin).
Please sign in to rate this answer.
Vladimir Mushkin 6 Reputation points

2021-05-05T17:45:19.17+00:00

Hi. I got some scanned documents and all the structure is OK but they are in Russian... So the output is a mix of a latin and cyrillic characters. This is very sad for me as I found this service really incredible for my purposes. I can easily parse non-scanned - programatically created PDF documents without cloud services, but I have this problem with images. Is there any kind of solution? Can you suggest anything? Maybe I can use some options for my C# client?

Ramr-msft 17,621 Reputation points

2021-05-28T05:23:15.42+00:00

@Vladimir Mushkin Thanks for the update. The Computer Vision Read API is Azure's latest OCR technology (learn what's new) that extracts printed text (in several languages), handwritten text (English only), digits, and currency symbols from images and multi-page PDF documents.
Sign in to comment

Share via

Form Recognizer cyrillic text in scanned document

1 answer