<big_grin>
I think that first you should Ask the Dark Gods. Certanly that will help.
Make sure to ask them exactly what you've asked us: Take a number from you, convert it to an OCR font, put it in a PDF file, print it out, run the printout through an OCR scanner, and give your number back to you.
If the Dark Gods can't help, purchase some good OCR software and PDF authoring software. It's too much work to reinvent that wheel from scratch.
</big_grin>
Seriously though, the problems are that your specification is unclear, and Perl isn't the best tool for this particular job. OCR interpretation is a "Big Job", and most of the companies that specialize in it charge for their software. ...and I've yet to see flawless OCR that can handle more than a very limited type of input.
Some suggestions have been offered. But honestly, I think you're probably either going about it the wrong way, or trying to crack too big of a nut, or both. If serious OCR conversion is your goal, you're better off paying a couple hundred dollars for a well-developed tool. As mentioned before, the fact that there is virtually nothing on CPAN for OCR (and yet OCR is such a well-known task), should be a strong indicator that in this case, home-made Perl scripting isn't your solution. But maybe I've simply not understood what you're trying to accomplish.
Update: <big_grin> tags added. |