Hello Manuel. Thanks for your interest in ocrad.
Manuel Mall wrote:
In our application we provide our users with on-line coversheets they use to
send faxes back to us which we scan. Obviously we have control of the
layout, fonts, sizes etc. of the coversheet. As we would like to use the
scanned results of the faxed back coversheet to automatically associate the
fax with a customer/order we are wondering which fonts and font sizes have
shown to give the best results when performing OCR with OCRAD.
I have never tried ocrad with faxed texts, and I haven't got a fax machine to play with (nor time to do it). But, as the README file states, "For better results the characters should be at least 20 pixels high". So, perhaps a lowercase height of 20 fax pixels will work well.
In any case, you may try different sizes and send me a couple of the best recognized images (in pbm format), so that I can improve ocrad's recognition of fax images and include some info in the documentation for future users.
Regards, Antonio Diaz.
_______________________________________________ Bug-ocrad mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-ocrad
