3.968 OCR software: TextPert and OmniPage (96)

Willard McCarty (MCCARTY@vm.epas.utoronto.ca)
Mon, 29 Jan 90 19:45:31 EST

Humanist Discussion Group, Vol. 3, No. 968. Monday, 29 Jan 1990.


(1) Date: Mon, 29 Jan 90 12:34:47 EST (17 lines)
From: cbf@faulhaber.Berkeley.EDU (Charles Faulhaber)
Subject: Re: 3.960 OCR software? French e-dictionary? (49)

(2) Date: Mon, 29 Jan 90 15:29:37 EST (7 lines)
From: "Michael S. Hart" <HART@UIUCVMD>
Subject: Re: 3.965 scanning software (22)

(3) Date: Mon, 29 Jan 90 11:28:39 EST (18 lines)
From: David.A.Bantz@mac.dartmouth.edu
Subject: Re: 3.960 OCR software? French e-dictionary? (49)

(4) Date: Mon, 29 Jan 90 16:11:13 -0800 (35 lines)
From: Malcolm Brown <mbb@jessica.Stanford.EDU>
Subject: OCR software: don't crown TextPert yet!

(1) --------------------------------------------------------------------
Date: Mon, 29 Jan 90 12:34:47 EST
From: cbf@faulhaber.Berkeley.EDU (Charles Faulhaber)
Subject: Re: 3.960 OCR software? French e-dictionary? (49)

We are using OmniPage (Caere Corporation, Mountain View, CA) with a
Mac IIcx and Apple Scanner. There is a foreign language version which
will handle most (all?) of the modern European languages with Roman
character sets. I don't know about Greek, but I doubt that it can do it.
We have been using it with Spanish. It handles straight text quite well
if the copy is good. It was not very good with a critical edition of
18th-c. poetry with a complicated apparatus.

So far we've only run some tests, but we will be going into production mode
this semester.

Charles Faulhaber
UC Berkeley
(2) --------------------------------------------------------------15----
Date: Mon, 29 Jan 90 15:29:37 EST
From: "Michael S. Hart" <HART@UIUCVMD>
Subject: Re: 3.965 scanning software (22)


I have researched both TextPert and OmniPage, and found OmniPage the
more flexible. mh
(3) --------------------------------------------------------------29----
Date: Mon, 29 Jan 90 11:28:39 EST
From: David.A.Bantz@mac.dartmouth.edu
Subject: Re: 3.960 OCR software? French e-dictionary? (49)

Omni Page recognizes diacritics but is not trainable. AccuText (from
Xerox) we find to be generally more accurate, but it does not recognize
diacritics (that is promised in the future). TextPert is *supposed* to
be able to deal with Greek and Russian as well as roman-based languages,
but there is no working demo from them, and we haven't gotten the
courage to send in the $800 or so to purchase it.

--- Michael W Jennings <MWJENNIN@PUCC> wrote:

Does anyone have experience with or know of OCR software that will run on
a Mac IIcx with an Apple Scanner; we specifically need software that
recognizes or can be trained to recognize foreign languages (German and
Greek).
--- end of quoted material ---
(4) --------------------------------------------------------------46----
Date: Mon, 29 Jan 90 16:11:13 -0800
From: Malcolm Brown <mbb@jessica.Stanford.EDU>
Subject: OCR software: don't crown TextPert yet!


I'm currently working on an article that will be a survey of Mac OCR
programs. Much to my surprise, the results I have obtained with the
latest release of OmniPage have been *very* impressive. I haven't run
OmniPage through all my tests, but the results so far have been quite
good, outdoing TextPert and even the Kurzweil 4000. I'm working with
OmniPage 2.1. Indeed, one of the big objections to Omnipage, from the
scholar's point of view, has been removed: the program now supports a
variety of latin-based character sets. One of my tests involved
scanning some photocopied pages of "Also sprach Zarathustra." Behold!
not only did OmniPage get the umlauts correct, it distinguished
correctly between majuscule U and majuscule U with an umlaut, something
other programs seem to always get wrong. Moreover, it scans pages in
half the time of TextPert (at least when using the HP ScanJet+, as I am
doing) The primary disadvantage of OmniPage is that it requires so much
RAM. Just thought I'd muddy the waters a bit --- also begann der
Untergang...

Malcolm Brown
Stanford