[Insight-developers] Use physical 0, 0, 0 instead of voxels' center

Vivien Delmon vivien.delmon at creatis.insa-lyon.fr
Mon May 21 08:50:11 EDT 2012


On 05/18/2012 04:30 PM, Miller, James V (GE Global Research) wrote:
> Vivien,
>
> We are still looking at your patch. I need to dig into what coordinate
> frames the various calculations are being done.
>
> However, with a change like this where the output of a test needs to
> change, it is best to correct the test. This may simply require
> submitting a new baseline image. This is a two step process as the
> baseline images are maintained using a content based hashing system. The
> input file in the repository is just a small text file containing the
> md5 hash for the new image. That file needs to be updated. Then the new
> image needs to be uploaded to the system which serves the baseline images.

I pushed the image resulting from the test. It seems to be the one 
expected at first, since in this new image the projected cube is centered.

> I "think" this can all be done at the gerrit-push stage. Just put the
> updated baseline image where it needs to go and the build process and
> gerrit-push process should create the md5 version of the file and push
> the new data to the server.

My push created a new topic in gerrit 
(http://review.source.kitware.com/#/c/5841/), I don't now if it's the 
normal. I had to add my reviewer again so I certainly missed something 
during the process ...

> Here is link that describes the process
>
> http://www.itk.org/Wiki/ITK/Git/Develop/Data
> Jim
>
> On May 16, 2012, at 4:24 AM, Vivien Delmon wrote:
>
>> Hi,
>>
>> I pushed my first path into guerrit here:
>> http://review.source.kitware.com/#/c/5639/
>>
>> Hans J. Johnsonrequested for a build in which a test fail which is
>> normal since my patch changes the geometry used by the
>> RayCastInterpolator. What should I do now ? correct this test to take
>> into account the physical coordinate system of the image and submit a
>> change ?

-- 
Vivien Delmon
PhD student at CREATIS, Lyon
Supported by ANRT and ELEKTA
http://www.creatis.insa-lyon.fr/rio



More information about the Insight-developers mailing list