[Debian-hebrew-package] HOCR is not scanning properly
kobi.zamir at gmail.com
Thu Apr 9 05:14:49 UTC 2009
thank you for your bug report and attached image.
1. current version of hocr (0.10.x and also 0.8.2) can not recognize numbers :-(
2. hocr use statistics about font shape and size. if your image has
less then 10-20 letters it will not be able to seperate letters like:
י ר ך ו
a. use texts without numbers.
b. use texts with lots of letters
3. you can use the debian hebrew mailing lists.
>Good day, Yaacov!
>Since I've not found any mailing for for hocr, I'm writing to you directly.
>I'm using hocr-0.8.2 (Debian Lenny) and trying to ORC simple image
>with Hebrew text and number (attached).
>Unfortunately hocr-gtk fails to process it - it recognizes only "lamed".
>Am I doing something wrong?
More information about the Debian-hebrew-package