From cvance@cfar.umd.edu Sun Jul  3 12:28:17 1994
From: cvance@cfar.umd.edu (Christopher Vance)
Newsgroups: alt.sources
Subject: Re: OCR sources?
Date: 27 Jun 1994 19:58:11 GMT
Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742
NNTP-Posting-Host: empedocles.cfar.umd.edu
X-Newsreader: TIN [version 1.2 PL2]

Joerg Behrend (aba09@rrz.uni-koeln.de) wrote:
: To learn something about OCR, I am looking for such a program in source code
: (C/C++ preferred, but other standard languages are also welcome). Also parts
: of an OCR program (like segmentation of characters or assignment of an
: unknown bitmap to a character) are of interest.
: Thanks in advance.

: Joerg Behrend,  aba09@rrz.uni-koeln.de, I. Physikalisches Institut Uni Koeln


	The only public domain OCR software (for unix) that I have found
was written by some students at CalPoly, you can get the software by
ftping to polyslo.csc.calpoly:/pub/ocr
	Based on that software is an improved version by Jason Molenda,
available at ftp.cygnus.com:/pub/gnucr-0.0.tar.gz

If there is additional code available, I would be interested in hearing
about it, and would appreciate it if you could send me some more information.
OCR is one of the projects I have been working with for a while now.

I have spent a large amount of time going over that software, re-writing
just about everything except the neural net.  Please contact me, and
let me know more about what you are looking for.  I am posting here to 
let everyone else know that such software is available, and unfortuantely,
right now, it doesn't work all that well, but people are working to 
improve that.  The software I have re-written isn't ready for any 
sort of release (it is still a mess) but the with the improved 
segmentation routines, it works much better than the original software.

I have also set up a WWW server for the Document Image Understanding
group for the Center for Automation Research (CfAR) at the University
of MD, College Park.  That site is accessible at 
http://documents.cfar.umd.edu/  It is currently a repository for
Document Image Understanding and Character Recognition information and
resources.  You can check there for more information.
	I had to put in a plug for the site :)