Simple and reliable method to preform an OCR on a pdf? [SOLVED]

graysky · 2014-03-30 09:51:54

I have Acrobat at work which does a wonderful job preforming an OCR (optical character recognition) on scanned pdf files so that one can do keyword searching within the pdf. I have spent the past hour following-up various guides on the 'net at large which do not work for me. Can someone recommend a simple and reliable method to accomplish this?

Last edited by graysky (2014-03-30 13:50:34)

korn36 · 2014-03-30 10:47:35

OCRFeeder maybe? I never tried it, but it looks like it will do what you want it to do.
It supports multiple OCR engines such as Cuneiform and Tesseract.

Last edited by korn36 (2014-03-30 10:52:26)

jsoy9pQbYVNu5nfU · 2014-03-30 13:13:45

I can second OCRFeeder. It can use different OCR engines, and the application doesn't hide the complexity of the options that these engines accept, so with a little fine tuning you should get reasonable results. For Latin text, that is. For non-Latin or mixed scripts, all FOSS engines are essentially useless (I've tried long and hard to get Japanese OCR to work).

To be honest, I myself only use Acrobat now because it's the only thing that produces consistently excellent results. For example, text selection will often not work smoothly because the FOSS tools don't group the text objects properly into lines or paragraphs.

graysky · 2014-03-30 13:50:14

Thanks for the suggestions and the tips. I will give that a shot and compare the results to acrobat pro which I am licensed to use at work. I will say that acrobat is very trivial to use even in batch mode and gives high quality results in my experience. Which engine do you recommended for US_English sources?

Optional dependencies for ocrfeeder
    tesseract: character recognition
    ocrad: character recognition
    gocr: character recognition
    cuneiform: character recognition

korn36 · 2014-03-30 14:20:13

Try tesseract, I've used it once for OCR'ing some text and it worked fairly well. Don't forget to install tesseract-data.

Arch Linux

#1 2014-03-30 09:51:54

Simple and reliable method to preform an OCR on a pdf? [SOLVED]

#2 2014-03-30 10:47:35

Re: Simple and reliable method to preform an OCR on a pdf? [SOLVED]

#3 2014-03-30 13:13:45

Re: Simple and reliable method to preform an OCR on a pdf? [SOLVED]

#4 2014-03-30 13:50:14

Re: Simple and reliable method to preform an OCR on a pdf? [SOLVED]

#5 2014-03-30 14:20:13

Re: Simple and reliable method to preform an OCR on a pdf? [SOLVED]

Board footer