It would be really cool if I could integrate unpaper post-processing
with pct-scanner-scripts!

Thanks,

also sprach Jelle de Jong <jelledejong@powercraft.nl> [2011.07.08.2002 +0200]:

I'd suggest --script-dir too.

What would be input and output of this script?

Sure, although a tighter integration of unpaper would probably
benefit more users and require less local customisation. It's a nice
piece of software.

Yes, I have not been going anywhere near OCR yet. Btw, it would be
good for PDF too, not just DjVu.

Thank you for the report! Your request for unpaper is valid, I see
enough cases where this may be usable.

I have been thinking about a way for users to run there own scripts

I may add an --script ... options that allows a path to a script or
application that can run on the page.

That will be called in the processing script as first step before the
image will be used. The input image can then be changed by this
"script/app" as the users wishes...

Will this fill your needs?

I have also been looking at OCR for the DJVU output, but have not yet
succeeded in usable OCR results in Dutch or English with floss software...

Kind regards,

Jelle