1998.06.03 15:01 "Re: Zones", by Helge Blischke
I have a need for a function that can analyse TIFF images, which will be of scanned periodicals (Newspapers, Magazines, Newsletters etc.) I need to look for whitespace as a guide to creating zones (OMNIPage 7 has this facility - but is very limited in quantity (64) and anything more complex is rejected immediately.) To explain further, these files are scanned on one side of the world and sent electronically to a _very_ large data warehouse processor which would dissect each file. (The zones containing text would then be OCR'd and the graphics would then be many pages of one new TIFF file.)
The only problem with my request is that I am... a VB programmer, and not too conversant with C, beyond DLL declarations and all that... sorry.
As far as I know (but didn't test it yet) there are at least 2 OCR engines that promise to do what you want beyond the limited capabilities of OMNIPage:
FineReader (see: http://www.abbyy.ru)
PrimeOCR (see: http://www.primerec.com)
You may download a free demo version of FineReader; with PrimeOCR, you can submit test images they'll process for you.
Whereas PrimeOCR is rather expensive (sold only as an OCR server application on NT), FineReader is fairly cheap (maybe even free for personal use only).
If one of these products (or some other I don't know about) matches your needs, I think it's better to use it but to hack the image interpretation yourselves (think of skewed, slightly rotated or distorted images).
Hope this helps (a little bit at least)