You are not logged in.

#1 2012-05-12 14:51:02

rnabioullin
Member
Registered: 2010-10-02
Posts: 24

Suggestions for document scanning program/workflow

I am in need of a program/workflow for scanning a variety of mixed text/image media, including books, government and financial documents, university assignments, etc., which will at minimum allow one to crop and adjust the bitonal threshold, and produce djvubind files with OCR data. I have originally used a custom bash script to implement a scanimage - GIMP - djvubind + tesseract workflow, but it was inconvenient and djvubind was buggy in integrating tesseract (despite sending a bug report that was resolved). I also tried scanimage - scantailor/scantailor-cli - djvubind + tesseract, but scantailor had editing limitations and of course the OCR bugs were still an issue. My latest approach is gscan2pdf, a program that does properly integrate with djvulibre and tesseract, but the program is very buggy and therefore painful to use on a regular basis - for example, I have to manually verify every page twice (since sometimes it does not invoke OCR on certain pages, and sometimes it does not export all pages to the djvu file).

Ideally, it would probably be possible to achieve the benefits of both automation and a plethora of editing capabilities, while reducing redundancy, by some sort of GIMP plugin/extension. Any ideas?

Offline

Board footer

Powered by FluxBB