Slashdot Mirror

← Back to Stories (view on slashdot.org)

Open Source OCR Packages?

Posted by Cliff on from the "o-c-r"-or-"o-c-a" dept.

MoonCat asks: "Are there any Open Sourced OCR (Optical Character Recognition) packages available which are comparable to their commercial counterparts in terms of recognition performance? "

11 comments

  1. No Such thing ... at least not alive ... by Zeljko+Blace · · Score: 1

    there was one project in New Zealand(they even offered a job for that but no one took it ... great shame ... one of those holes in open source apps ... just like VIDEO editing and sound sequencing .

  2. Qui-ne-faut by Ray+Dassen · · Score: 2

    There's Qui-ne-faut. I haven't studied it in detail, but on a trial run it was at least isolating character shapes well (though the actual identification of characters from them left much to be desired - probably there's some training needed).

  3. How about payware? by spinkham · · Score: 1

    Are there any "payware" OCR programs for linux either? I have a Linux supported scanner, and would like this functionality, free or not..
    Of course, Open Source is prefered, but right now I don't know much about the field, so couldn't hack on it for my self very much anyway.

    --
    Blessed are the pessimists, for they have made backups.
    1. Re:How about payware? by bp · · Score: 1

      Recognita has an unsupported Linux port. Ask for it, the more people asks, the better :-)'

      --
      -- what is a sig?
    2. Re:How about payware? by bruceg · · Score: 1

      I use OCRShop from Vividata. It works very well, and their Linux pricing is less expensive than the commerical UNIX versions. Their OCR engine is derived from Caere OmniPage. The Linux version only outputs text, but converting to other formats is pretty simple, especially if you have StarOffice or Applixware.

      http://www.vividata.com

      They have just upgraded their OCRShop for Linux to version 4.5, and it is compatible with RedHat 6.0 (So I'm assuming it would probably work on any GLIBC 2.1 based system) Good luck!

  4. I don't know, but someone does... by Pauly · · Score: 2

    Check out the project that reads all of our
    USPS mail:

    http://www.pnwsoft.com/linux.htm

    I imagine John Taves, the consultant behind this
    masterpiece, could tell you a thing or two about
    OCR on Linux. His email address is at the bottom
    of the article.

    Having worked for years on OCR, trust me when I
    say the success of this project is a minor miracle!

  5. OCR position, that was me! :-) by Stuart · · Score: 2

    Hi all,

    The department of Computer Science at Waikato University in New Zealand is offering the position. It's a great place to live, they pay well and you get to write GPL'd software. They even pay to fly you there!

    If you want more information write to Professor Ian Witten and ask about the GNU OCR job! It's a University so they're not after people with 10's of years experience, they're after hackers that can get stuff done :-)

    1. Re:OCR position, that was me! :-) by ansible · · Score: 1

      Waikato is working on many interesting things, including an OOP called Sather (also a GNU project). I would love to fly down there to work on my Master's for a couple years.

      But as with many things, I just don't have the time. I suppose I need to work harder so that I can retire early.

    2. Re:OCR position, that was me! :-) by Anonymous Coward · · Score: 0

      A lot of people who visit Waikato, or New Zealand in general, find their lives enriched. Not just professionally but spiritually (for whatever that means to each of us) as well.

      For the hackers, N.Z. has better internet connectivity than anywhere in the world, with the capital (Wellington) being the "most wired" city in the world (according to Wired). If you want to be on the bleeding edge, live a life of enjoyment and love the outdoors you should visit here ASAP :-)

  6. Re:Free OCR??? (Check NIST) by Anonymous Coward · · Score: 0

    The National Institute of Science and Technology (NIST) has been developing OCR software for some time now (mainly for reading census forms as I understand it). Anyways, since it's a public agency, the code and data they produce are entirely public domain. (At least in the US, I don't know if there are any issues outside the country) You can actually write them and they will send you a set of CDs with the source, sample data, and printed instructions on compiling and installing it. Last I knew it would build on U*IX, but looked like an awful lot of porting to go anywhere else. Oh yeah, a quick search using "NIST+OCR" turned up the site on yahoo. Good luck.

  7. OminiPage Works in my office here... by Anonymous Coward · · Score: 0

    Granted, it is under Windog, but if the engine is the same, it should work beautifully.

    John Gunkel