ocr success

[Krishnakant]

hi
daniel,
I think this is one of the greatest achievements.
A few days back some hacker on the orca list had sent me a script for
doing ocr.
I have some how misplaced it i believe and I would request for another
copy.
ah, willem was that you who sent me the script?  if yes kindly send it
again.
I have one major question?  will any of the tools mentioned in the
original post do rotation of the image if I place the page up side down?
I think most free ocr softwares can't do it but I will be pleased to
know that I am wrong in my analysis.
I think we really need a howto from daniel for this.

happy hacking.
Krishnakant.
On Sat, 2008-12-20 at 20:42 +1100, Daniel Dalton wrote

> Hi All,
> 
> First sorry if you get this a couple of times, I thought I would just
> share my experiences with everyone, but if it's not wanted, I won't do
> this next time...
> 
> Anyway, I've had great success with ocr on linux.
> Using the following packages:
> scanimage: - grab the image from the scanner
> - kies: Front end for the ocr and scanimage stuff
> - Teseract: ocr software 
> - ocropus: ocr software I believe 
> 
> What I did:
> Install kies.
> next, checkout tesseract from svn and build:
> patch -p0 < the_attach-patch.diff
> ./configure
> make
> sudo make install
> Note: the patch is required to make it compile.
> Next checkout ocropus (r864) from memory.
> ./configure
> jam
> sudo jam install
> to compile/install it.
> Next run
> sudo chmod %s /location/to/scanimage
> (note install sane from apt)
> sudo scanimage -L
> if it finds your scanner do this:
> scanimage > imagefile
> If it scans it as an image, try the kies front end:
> kies_p2t
> Press space and wait for your document to open in lynx.
> To save it press p from lynx and save it in the usual way...
> 
> Hopefully I didn't miss too much, and I know it's very brief. Let me know
> if there is interest in this subject, I'll place a howto on my
> website... I'm also going to (if its ok with Willem), generate a diff to
> improve the documentation slightly although it's pretty good at the
> minute, and the scanimage command he's script uses needed to be modified
> on my system...
> 
> So is there interest in a howto for ocr on linux?
> 
> Let me know if you have any questions. I would also like to thank everyone
> involved with all the ocr projects, and also sane/scanimage.
> And I would also like to thank Willem for all his great help off list!
> 
> So if there are any questions about this or you would like me to write
> up a howto let me know...
> 
> Cheers,
> 
> Daniel
> 
> _______________________________________________
> Blinux-list mailing list
> Blinux-list at redhat.com
> https://www.redhat.com/mailman/listinfo/blinux-list