read multi-language ( arabic and english) image

[thinker]

reading image  with multiple language (arabic and english)  by using  -l ara+eng option gives garbage output.

how to read that image with both arabic and english content?

[Nick White]

On Fri, Jun 27, 2014 at 01:48:52AM -0700, thinker wrote:
> reading image with multiple language (arabic and english) by using -l
> ara+eng option gives garbage output.

There are currently a couple of bugs with combining Arabic and
English together, so it isn't working. I'd recommend you add any
extra information you have to those bugs, to help the issues be
resolved sooner:

https://code.google.com/p/tesseract-ocr/issues/detail?id=899
https://code.google.com/p/tesseract-ocr/issues/detail?id=1220

Nick

[Janusz S. Bien]

Quote/Cytat - Nick White <nick....@durham.ac.uk> (Fri 27 Jun 2014
06:49:46 PM CEST):

In the meantime you can try to merge the results of the separate runs
for each language. You will find hocr-merge at

https://bitbucket.org/jwilk/marasca-wbl

in misc/xhocr /

Best regards

Janusz
--
Prof. dr hab. Janusz S. Bień - Uniwersytet Warszawski (Katedra
Lingwistyki Formalnej)
Prof. Janusz S. Bień - University of Warsaw (Formal Linguistics Department)
jsb...@uw.edu.pl, jsb...@mimuw.edu.pl, http://fleksem.klf.uw.edu.pl/~jsbien/

[thinker]

thanks for your reply.