Re: OCR In Linux

["Martin G. McCormick" <martin@dc.cis.okstate.edu>]

	Since this topic has come up, I have a few questions.  Do
any of the free OCR packages run in command line mode?

	I don't mind that they may not automatically turn the
image for you to take care of upside down pages, etc, but how
does the quality of the conversion compare with something like
OsCar or Reading Edge?  I have found those products to do a
pretty fair job on standard printed pages or typewritten
material.  If I could duplicate that functionality in Linux, I
would be happy as the OsCar system I use here at work is
approaching 9 years old and the HP scanner will probably crash
and burn one of these days.

	It is my understanding that you do the OCR under UNIX by
scanning a Jpeg image with the scanner and then feed that image
in to the OCR program where you get a text file or standard
output if text can be decoded.

	If the output indicates differently when tables or
columns are involved, one should be able to write filters in C or
perl or even awk to read them more easily.

Martin McCormick WB5AGZ  Stillwater, OK 
OSU Center for Computing and Information Services Network Operations Group