PDA

View Full Version : OCR Program That Works...



zandoval
09-05-2005, 08:13 AM
The following works with documents that have already been scanned - xsane is what I use to scan and getting my scanner to scan was a real ordeal... But assuming that you are succesfully scanning this OCR program is for the cost - THE BEST...

I found and am running a great OCR program - Much better recognition than anything in kooka or xsane and the other OCR programs I tryed - Its a windows binary that runs well in wine - All it does is the OCR so you have to scan your documents first and save in .tif .tiff bmp. or .gif - After you scan the documents you can find the program in your .wine file, open it and single or batch the scans for OCR - It does a great job even at 400 Mhz with 128 meg - It might not work trying to run on some slower computers but the programing is very clean and simple AND ACCURATE...

Its a windows freeware program found here... "TOCR"

http://www.transym.com

This guy deserves a lot of credit for making a clean enough program that it can run well in our linux systems...

Currently running in KANOTIX 3.9 HD install - Also works in KANOTIX Lite HD Install - Feather Linux HD install - And of course BIG DADDY KNOPPIX 3.9

Notes: I down loaded the zip file and expanded it - Then ran the install under wine - Then found the viewer.exe at file:/home/yourname/.wine/drive_c/Program Files/transym/ - On opening the program I find the file I want to OCR - I use xsane to scan my documents and I save them in non compressed .tif files - On one of the computers I set this program up on I had to reopen the scans in gimp then resave them due to some auto file compression problem with the scanner - gimp could open the files but tocr could not untill I opened them in gimp and resaved them.

There is another OCR program I have run in wine called "SimpleOCR" it has more bells to it but just does not run as good an OCR as TOCR.

Let me know what you all think - OUT

Harry Kuhman
09-05-2005, 10:06 AM
I've loked over the information on their website, but I really can't be bothered to download and try fooling with this software. Is your tip really to get a $40 piece of downloaded OCR software that is intended to run under Windows and run it under wine? I've used several OCR programs and they seem much more feature rich than this, and they were included with my scanner or cost the same or less than this. You may call features "bells and whistles", but I find things like following text as it flows from column to column and being able to OCR even when the text wraps around graphics to be pretty useful. And I have absolutely no complaints at all about the quality of the recognition I have gotten from the software that I have used. There are even "bells" in that area that this simple program seems to lack. For example good OCR software will support a high quality check of the final text to try to catch potential errors, as well as an editor that lets you see the final text and the part of the scanned image that it comes from side by side. As I say, I seldom have any OCR accuracy issues, but one time when I did OCR many pages of typewriten text, with x-outs and strikeouts with hand printed words squeezed over the bad word as a correction, such a "bell" made the work a breeze. I don't want a simple program that lacks this.

I have done all my OCR work in Windows. This too is a Windows package. Have you compared it to any Linux OCR software, such as clara? I would be interested in seeing some hard figures that support the claim that this is better, and not just simpler and more expensive software.

Just out of interest, are you in any way associated with the production of this software?

zandoval
09-06-2005, 05:24 PM
Harry - Thank you for your reply - I am in no way conected with the software manufacturer - I am a guy useing nothing but knoppix/kanotix type OSs on his old beat up computers - I tryed for two months to get both scanners and OCR programs working in these sytems - I was surprised how hard it was but that is complicated by my limitations of Linux in general - That is why I use HD installs of live CDs - The OCR programs that came with my scanners did not work in linux - Even the HP scanner programs - And HP has been pretty easy to run in linux - At my original post I was not aware that TOCR was in fact shareware - It costs 49.00 US - For me this is to much as I am a real cheap guy - But still - It is a program that runs in wine rather well and certainly deserves some credit - I could not get CLARA working on my systems - But this is probably my limitations as a user - GOCR used in xsane and KOOKA work - But are not as accurate as TOCR - I needed accuracy in an OCR program becuase I was trying to OCR a bad copy of some field manuels that were about 10 to 8 font in size - Even when scanned greater that 600 most OCR programs failed - But TOCR worked - I have not tryed any of the LINUX commercial products - I have used windows OCR programs on windows computers and yes they do spelling checks, collums, edges, and pics, plus save into many formats, but allas.... I do not have a single windows computer in this house of six - Please let me know if you have any sugestions on how I can better OCR with my limitations - I have three more field manuels to OCR - Two of them water damaged...

Harry Kuhman
09-06-2005, 07:09 PM
Sorry, no suggestions. I havn't done OCR in so long that I don't remember the name of the last program I used. Everything has been in Windows (although the above program is a Windows program as well). And the price sure did put me off too for what was offered. No sense for me even looking at crippleware to evaluate it when I know I will not send him all he wants and will have to remove it later. I havn't used any Linux OCR software, as I still run both OS's, my scanner works under Windows, and as I said I have had good OCR software included in the price of some scanners that I have bought. Hopfully others will join in with more information.