Skip to main content

For blind Linux users who sometimes need to use OCR to get some kind of access to screenshots, VM screens, just plain inaccessible UIs, etc., how do open-source OCR engines (e.g. Tesseract) compare in accuracy, particularly for screen images, to the Windows OCR engine that NVDA's OCR feature uses?
in reply to Matt Campbell

Peter Vágner
Peter Vágner
•
I'm using #ocrDesktop that uses #tesseract under the hood and it works fine. Of course there are some inaccuracies here and there but for checking VM status, reading teamviewer password and similar tasks it's enough. Here is a wiki article about it. https://wiki.archlinux.org/title/Ocrdesktop
#tesseract #ocrDesktop
⇧