| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 4483.1 | there's no easy answer (I don't think) | NOTAPC::SEGER | This space intentionally left blank | Thu Mar 14 1996 13:11 | 30 |
| I think I may need a little more info. For example, as far as I know, the
hardware and the software are two different things. Most scanner vendors
package someone else's OCR software with their hardware.
If you're looking to do full sheets of paper, you really need a flatbed scanner.
There may be some with sheet feeders that will HAVE to be software controlled,
since you need to be able to SCAN - OCR - SCAN - ... and the pause in between
scanning and OCR'ing can be a minute or more depending on the detail.
Most scanning packages I've seen scan text into a work processing package like
MS-Word. It sounds like you're saying you have preformatted forms. I don't
know if there are any products out there that do this or not. I suppose if you
had to you could scan the forms into a word processing document and then write
a macro that reformats the data in a form suitable for loading into a database.
Putting it all together in an automated fashion is yet another issue. Is this
an absolute MUST? DO you have 100's of documents? 1000's?
and then of course there's your desire for something cheap. I'd think there
would be a direct relationship between the level of automation and the $$$,
hence my curiosity of knowing how many documents are involved. from what I've
seen, flatbed scanners (without automatic feeders) tend to sell for about
$500-$1000 (and of course more). I've seen OCR software ranging from as little
as $100 to over $1000.
oh yeah, one more thing - no OCR software does a 100% perfect job, so even if
you could automate the entire thing, you'd need to manually go back and check
each item in your database to make sure it was scanned correctly.
-mark
|
| 4483.2 | | DECWET::VOBA | | Thu Mar 14 1996 15:49 | 8 |
| Re .0, you'd need to grab a copy of Imaging World (IW) monthly trade
tabloid. It's one of the many publications dedicated to this area of
technologies. If nothing else, the ads in there would give you plenty
of pointers to a myriad of vendors who are providing solutions and
technologies that you are seeking. The paper can be reached via
[email protected] or http://www.btb.com/cardinal/iw.
--svb
|
| 4483.3 | A working configuration. | MUGGER::BURKEP | Pete, Software Engineer, OSEC - UK SI | Fri Mar 15 1996 03:50 | 16 |
| I agree with the points made in .1, you need more information, and OCR is never
100%.
However, if you are after something that works:
I have a Fujitsu ScanPartner 10 flatbed scanner with sheet feeder, connected to
my Celebris 590 (W4W) via an Adaptec AHA 1542CP SCSI adapter and I'm using
Calerea Wordscan Plus V4.0 to control the scanner and to do the OCR.
I'm currently integrating this lot with LinkWorks for an SI project. My problem
is that I got what I could get. I didn't evaluate these products and I don't
know how these products stack up against any others.
Can anyone else recommend products in this area?
Pete.
|
| 4483.4 | Check after-sales support (Microtek is poor) | STAR::PIRULO::LEDERMAN | B. Z. Lederman | Fri Mar 15 1996 08:09 | 3 |
| Something else to check is after-sales support. I can state from
experience that Microtek is very poor. Other people posting in Usenet
News so as well.
|
| 4483.5 | AUSSIE::FAX | NPSS::GLASER | Steve Glaser DTN 2267212 LKG1-2/E10 (G17) | Fri Mar 15 1996 09:03 | 3 |
| AUSSIE::FAX has lots of data, not sure what's relevant.
Steveg
|
| 4483.6 | Flatbed scanners are *not* the only way | CSC32::T_BUNDRICK | | Fri Mar 15 1996 12:44 | 16 |
| Don't overlook the "little" scanners where the document is fed in from front -
like a fax machine. They can be a lot cheaper than a flatbed, and a
sheetfeeder is an option. I bought the *cheap* model for $299 at Sam's - does
up to 1200 dpi (no sheet feeder). The OCR included works well - a lot better
than the ones I've used with FAX s/w. This "intelligent OCR" learns new fonts
and you can multiple dictionaries. I've seen more expensive models at Couputer
City - mostly looks like same h/w with different label and different s/w
included.
I've pushed business cards and even newspaper clippings (cartoons, etc) through
mine and it hasn't ate one *yet*. I find the multiple dictionaries *very*
usefull... I usually name the dico file for the source - "jib" uses ibm
selectric typwriter ( clean, crisp font - not proportional), another user sends
me stuff from his pc - dot matrix, proportional, etc.
Tim
|
| 4483.7 | Thanks ... | TINCUP::MOE | | Mon Mar 18 1996 15:43 | 5 |
| Thanks for the replies ... I will check out the info supplied ...
assuming, of course, that this project is still funded since I asked
last week!
Brian
|
| 4483.8 | We have a Partner that does this! | MPOS02::BJAMES | I feel the need, the need for SPEED | Tue Mar 19 1996 18:50 | 11 |
| You should contact National Computer Systems, Inc. headquartered in
Minneapolis, MN. (612-829-3000) They build a wide array of scanning
solutions both OMR and OCR. Their home page is at http://www.ncs.com.
The product set they offer is called Optrieve and is a COLD based system
whereby it writes the scanned info. out to optical disk. If you need
further info. drop me an e-mail.
Regards,
Bill James
NCS Account Sales Manager
|
| 4483.9 | | SETIMC::OSTMAN | http://setimc.soo.dec.com/ostman.html | Thu Mar 21 1996 01:53 | 5 |
|
Have you considered Omtool Image Express (earlier known as DEC Image
Express)? You can find out more at http://www.omtool.com.
/Kjell
|