[sword-devel] Alternate Versification question
thulester at gmail.com
Fri Aug 10 06:36:24 MST 2012
Yes from the CLI (initial training is painful - though skillful Internet
searches can often get you training files for particular fonts that are
close). I tried using Qtesseract, but the compilation was somewhat
problematic, and felt the time investment (to get the thing to work) was
little worth the return (and went back to the command line).
Besides, if you get Tesseract trained well to work from the command line
you can automate most things in typical unix fashion via scripts and pipes.
Because I'm mostly lazy, what effort I do spend, is spent making things
work more automatically. Now that Tesseract takes .png files directly,
messy conversion business to .tiffs is gone and so effort scanning now ends
up being mostly preparatory with some touch up thrown in (scanning errors
can often be corrected through intelligent use of ispell/mspell/aspell).
Even so, OCR work will always contain some degree of human effort.
Other than Qtesseract, I've no experience with other GUIs (I not so with
Python, for example, which cuts out most of the others)
On Friday, August 10, 2012, David Haslam wrote:
> Are you using Tesseract from the command line, or did you also install a
> http://en.wikipedia.org/wiki/Tesseract_%28software%29#User_interfaces GUI
> for Tesseract ?
> If so, which? and any comments worth sharing?
> View this message in context:
> Sent from the SWORD Dev mailing list archive at Nabble.com.
> Instructions to unsubscribe/change your settings at above page
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the sword-devel