0

First of all, I apologize if this is not the right place to ask this, but I couldn't think of anywhere else (maybe Stack Overflow?).

Anyway, I'm looking for a Optical Character Recognition software (OCR) to process my notes. The thing is that occasionally there is an equation there in the middle, so I was looking for a software that can process the text and the equations together that I can run in my Linux system.

Ultimately my goal is to create a LaTeX file from that, so it wouldn't hurt if the output was already in LaTeX, but I guess that would be asking too much.

I couldn't find anything online that did that, but I think that's mainly because I'm not using the right search terms (English is not my main language). I did find this question but it's from 4 years ago and I think this have changed since then.

If I could get one good software to process the text part of the notes, and another to process the equation part of the notes, I'd be able to put them all together already.

Does anybody know a way of doing this?

  • 1
    I agree with @Jaleks below. Running OCR on handwriting is a waste of time, unless you have the kind of handwriting which looks like print (extremely rare). OCR does a reasonable job on print, but that's about as far as it goes. And I don't think OCR can handle math at all. Bottom line - it's hopeless, imo. I'm not saying it's an insolvable problem, but the technology isn't going to get there any time soon either. Sorry. – Faheem Mitha Dec 18 '16 at 20:54
2

This isn't Linux specific, but http://myscript.com/ has a math handwriting recognition app that is half-decent. Not sure about OCR though.

  • That actually seems very good. Their more advanced tool is available only for apple and windows only, but they have an older tool that is available for android tablets, and I'm actually considering buying one just so I can use it instead of writing it in a notebook. – TomCho Dec 18 '16 at 23:24
  • 1
    You are also free to sign up for a developer account and use the handwriting recognition technology! This works on Linux-based systems! You use polymer or the like to host it locally – Hee Jin Oct 20 '18 at 21:37
1

My opinion: forget it. OCR programs on Linux are all (gocr, tesseract, cuneiform, ocrad) quite bad, even on scanned serif fonts, in my experience. You can completely forget it to recognize handwriting.
And I really will be happy if anyone proves me wrong.

Even on big commercial program for other platforms, like Finereader (good, as it allows to train badly recognized parts for getting better results in large scans and exports to more [binary] formats) or Acrobat (OK, but no features, just dumb, but OK, recognition) you will fail to get acceptable results with handwriting.
Also you will have no chance of recognizing math formulas with either of them, not even of scanned books.

(I know, that some tablets come with some handwriting recognition, but do not know any names of products for that personally…)

  • Thanks for the answer. I am checking into two solutions and one of them is this, which seems like it may do the trick, but I can't be sure with handwriting. Just thought I should give you a heads-up. – TomCho Dec 18 '16 at 23:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.