Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: Notesfiles $Revision: 1.7.0.10 $; site uiucuxc
Path: utzoo!watmath!clyde!cbosgd!ihnp4!inuxc!pur-ee!uiucdcs!uiucuxc!bantz
From: bantz@uiucuxc.CSO.UIUC.EDU
Newsgroups: net.micro.mac
Subject: Re: Devices for character recognition ?
Message-ID: <96900021@uiucuxc>
Date: Sat, 23-Nov-85 16:12:00 EST
Article-I.D.: uiucuxc.96900021
Posted: Sat Nov 23 16:12:00 1985
Date-Received: Sun, 24-Nov-85 11:12:53 EST
References: <235@daimi.UUCP>
Lines: 14
Nf-ID: #R:daimi.UUCP:235:uiucuxc:96900021:000:670
Nf-From: uiucuxc.CSO.UIUC.EDU!bantz    Nov 23 15:12:00 1985


Obviously digitizers like Thunderscan or Magic Eye aren't going to produce
text, even if they produced a good image at the character level, which, by the
way, they usually aren't capable of.

Optical scanners in the $few thousand range typically read a particular 
list of (typewriter) fonts - IBM selectric Courier 10, etc.

The only more general solution, to my knowledge, is the Kurzweil 
"intelligent" scanner, which can be trained to recognize any reasonable
printed or typewritten font, including multiple fonts on a page.

The Kurzweil will read -- not including training for the fonts -- close to
a page per minute, but will still require occasional correction.