On Sun, Oct 18, 2009 at 11:17:36PM +0530, Krishnakant wrote:
Thanks wil, for that information. Long time back I remember some one talking about a shell script which can do ocr and output into text.
Perhaps that could be me
I had requested the script and I am afraid I will have to request it again because I seem to have lost it.
Check out this little howto i put on my site about 6 months ago, perhaps longer http://members.iinet.net.au/~ddalton/projects/ocr/howto_configur e-ocr_linux.html Some notes - if you can clarify these that would be handy at some stage I'll make modifications where necessary: 1. My patch may no longer be needed. 2. If you can't compile tessaract then try rev 225. (idealy debug the problem depending on your programming knowledge:)) 3. Notify me of errors in the howto. 4. Double check you have willem's latest scripts. If you have a lot of time to spare, then do test out the latest code of both tessaract and ocropus and do email me any patches you do happen to write that fixes compile errors. Of course if you just want to get it working then as I said follow those instructions, and do try the latest tessaract, and if that won't compile, revert back to 225. I would be interested in hearing how you go the only insentive from my perspective to test out the latest code would be for ocr accuracy, but chances are that won't have improved significantly, but hopefully I'm wrong.
Or was it a set of scripts? I can't remember.
Yes.
The initial look at OCRFeeder gives me the impression that it should be somewhat accessible.
I should give it a go. I can't see the results being much better than my current set up as they both run through the same ocr engine, but maybe they are, I guess there is only one way to find out - compare results between the two alternatives.:)
In any case we can conver it into a project which I had proposed a few months back. Some thing like Open book or cursvel.
Collaborating all of this stuff into one good ocr system would be a worth while thing that could be done, but it seems it is just different front ends using the same libs... Dan
Attachment:
signature.asc
Description: Digital signature