Conversation

Notices

1. ilja (ilja@ilja.space)'s status on Sunday, 13-Feb-2022 20:28:09 UTC ilja

   in reply to

   • Amolith
   @amolith I have no experience with it, but there's an OCR bot on fedi and it uses https://github.com/tesseract-ocr/tesseract and it's docs say it also does pdf
   
   (bot repo is https://github.com/Lynnesbian/OCRbot/ )
   • Amolith (amolith@nixnet.social)'s status on Sunday, 13-Feb-2022 20:28:10 UTC Amolith
     Does anyone know good tools for applying OCR to a scanned document?
     
     I don't want any organisation, tagging, cloud, etc. features, just something akin to imagemagick that reads the PDF, does the OCR magick, then spits out the OCRed PDF to a separate file.
     Tagomago repeated this.
   • Dan Jones (danjones000@fedi.absturztau.be)'s status on Monday, 14-Feb-2022 08:28:13 UTC Dan Jones

     in reply to

     • Amolith

     @amolith

     I used to do this for a living.

     I built a cloud based document management system that would take scanned pages, OCR them, and store them as PDFs.

     We used Google Cloud Vision, which was overkill, but my CEO had a hard-on for Google.

     Tesseract should be all you need: https://guides.library.illinois.edu/c.php?g=347520&p=4121426

     Although, may I suggest using DjVu instead of PDF. DjVu is a better archival format. It’s much simpler usually results on smaller fine sizes. Many PDF viewers already support it. But I don’t know exactly what your use case is, so that may not be an option