Re: [tlug] Linux Kanji Optical Character Recognition (OCR) software?

[Matt Gushee <matt@example.com>]

Stephen J. Turnbull wrote:
> > > > > > "Dave" == Dave M G <Dave> writes:
>
>     Dave> But you never know. Can't hurt to just double check. Are
>     Dave> there any Linux options for kanji OCR?
>
> Heh.  I've been asking that for 9 years.  As of Dec 2005, the answer
> was that there was no decent open source OCR for linux period.  For
> many years the front of the pack was gOCR, but I never got decent
> results from it for purely numerical data---typing it myself was
> faster and more accurate.
>
> There are a few plausible candidates on freshmeat.net, including
> kognition (for KDE), libgocr, Clara OCR, Kadmos, and GNU ocrad.

A couple of years ago I scanned a book (don't worry, it's in the public 
domain ;-)); in preparation for that task I tested gOCR, Clara, and 
ocrad. I found that:

  * None of them were very good (at best, about as good as the
    commercial OCR program I was using on a Mac about 1998--forget
    the name).

  * Clara had a very strange interface and crashed a lot.

  * ocrad gave the best results, followed by gOCR. I didn't really get
    any results out of Clara, since I couldn't figure out how to use it.

And of course this was all in English.

--
Matt Gushee
: Bantam - lightweight file manager : matt.gushee.net/software/bantam/ :
: RASCL's A Simple Configuration Language :     matt.gushee.net/rascl/ :