OCR options

I have a colleague who has a lot of records scanning to do, and OCR is
required.

=?ISO-8859-1?Q?Erik_Richard_S=F8rensen?=...

,=20

=?ISO-8859-1?Q?Erik_Richard_S=F8rensen?=...

=2E=20

I have used OmniPage over the years and remains the preference. However,
the price has escalated substantially. The requirement is to run on mac.
OmniPage Pro X for mac is $1079 AUD ($499 USD).

=?ISO-8859-1?Q?Erik_Richard_S=F8rensen?=...

Glurb.:-(!! - That price has risen since my old version of Omnipage=20
Pro... And to that I wasn't too happe with it either on my systems.=20
There are too many problems with non-Latin fonts and substitutions...

=?ISO-8859-1?Q?Erik_Richard_S=F8rensen?=...

d=20

Steve Ball...

According to the Mac's calculator, at today's rates, $499 USD is $507.52
AUD.

Steve Ball...

Oops, make that $564.03 AUD.

Chris Brown...

quite, but:

AU site
OmniPage Official AU Site
Australia.Nuance.com

Malcolm...

There have been no updates, other than the price, since OSX 10.1 The
system requirements say:
"Mac OS 9.0 or later
Mac OS X (10.1)"

Note they do not say "10.1 or later". This is not an accidental
omission. It requires write access to files it installs in
"Applications" (a big security hole), but its OSX installers doesn't
set the permissions correctly on anything later than OSX 10.1. They
have not fixed this in the last five years. Their support section says
you can work around this by setting the installer to run in Classic (it
doesn't work), or boot into OS 9.2.2 to install (which does work if
your Mac is old enough to do it). Of course if you have an Intel Mac,
that is no help.

I bought a copy before the ridiculous price change (my G4 can boot OS 9
for installing it), but I would not recommend it now. It is
surprisingly good at converting TIFF to plain text, but the RTF output
often had bad formatting. Their support section is very helpful on
this: Fix the output yourself with a text editor.

Any opinions or suggestions on lower cost (or open source; haven't found
any) alternatives? It doesn't need to handle tables etc. just straight
character recognition.

=?ISO-8859-1?Q?Erik_Richard_S=F8rensen?=...

I've now switched to ReadIRIS Pro 11.5 and it is working beautiful and=20
fast on my OS X 10.4.x machines, and it's really easy to use. - And best =

of all - until now - there hasn't been any problems with non-Latin fonts.=

You can download a time limited demo version here:

cheers, Erik Richard

David M. Wood...

I would take a careful gander at tesseract (open source, being hosted by
google), which is available through the fink project and is QUITE good.

AES...

ReadIRIS was well regarded, and maybe cheaper? (Haven't looked
recently.)

=?ISO-8859-1?Q?Erik_Richard_S=F8rensen?=...

s=20

=?ISO-8859-1?Q?Erik_Richard_S=F8rensen?=...

on.

Hm, I don't find ViewScan rather good for text scanning - especially=20
not, if there are many various typefaces and inlined pictures as well.=20
VS is better for pictures only.

Here Silverfast is better. SF also are for varius brands of scanners=20
like Agfa, HP, Umax, Linotype, Nikon etc.etc... - I have tried the=20
Silverfast AI Studio package with an older Agfa scanner, and it worked=20
fine, and it works fairly good with my new HP scanner as well.

Cheers, Erik Richard

Malcolm...

this capability not mentioned on its web-page. They have a free trial
version.

Jolly Roger...

Well I can relay a recent experience to you. My grandfather purchased a
scanner last week for his Mac mini running 10.4, and wanted my
assistance setting it up since he's never done OCR before. The scanner
came with OmniPage SE (and it looked to be an older version). OmniPage
refused to work correctly, giving us "File not found" errors and the
like. After messing with it, and re-installing it a few times, we gave
up on it. Luckily, I remembered I had purchased a copy of Readiris Pro
a while back and never used it. We installed that on his computer and
it works great. I feel comfortable recommending Readiris Pro based on
this experience: