cuneiform team mailing list archive

["Jussi Pakkanen" <jpakkane@xxxxxxxxx>]

2008/10/6 Dmitry Polevoy <openocr.polevoy@xxxxxxxxx>:

> Please, translate to english following.

Google translator tool gives me the following:

> Need help in organizing the testing kernel recognition.
>
> One way of recognizing the control system is to compare
> the results of recognition for different versions (platform releases).
> We (Cognitive) there are a number of pictures and recognize the
> results (formatted txt) versions of win32. Ready to give in good
> hands 8) The first portion is in svn at
> http://devel.openocr.org/svn/openocr/trunk/cuneiform/stands or
> willing to look at the results of which linux-image version.
>
> Then you can think of the automation process.
>
> Thank you.
>
> PS: c 08.10.2008 to 20.10.2008 will not, then is ready to return to the
> theme of the organization testing the kernel of recognition.

In a nutshell, it seems that this is a test set produced by the
"smarttext" output format. This allows automatic performance testing
of the OCR kernel.

Making this work with the Linux version would be a very good thing
too. If someone wants to look into this, it would be great.
Unfortunately I won't have time to do it until after I have done the
next release. It should work on Windows too, using MinGW, so maybe we
could join the code branches back together with Cognitive and have one
single code base for all platforms.