[Insight-developers] MeshCompare test for the Insight Journal

[Luis Ibanez]

Hi John,

Good point...

Yeap, we don't have a MeshCompare,...
and as you correctly guessed,
the ImageCompare will not apply to Meshes.

It seems that we should add a MeshCompare in the future,
but it is probably to close for having this done before
the reviews for the MICCAI Workshop start.

Note that the difficulty with Mesh compare is that the ITK
mesh may have anything as type as PointData and as CellData.

We could start by making a Point coordinates comparison with
some tolerance for the average error, and maybe some evaluation
of the worst case (Haussdorff distance).

Note that you could do internal tests in your submission,
by adding a simplified version of a mesh comparison program
that you could include in your submission. Otherwise, you
may want just to submit a revision of your paper with an
additional paragraph explaining why the code in your submission
doesn't directly fit in the current testing framework.


Thanks a lot for pointing this


    Luis



----------------------------
jmelonak at ece.gatech.edu wrote:
> Hello,
> 
> In our recent submission to the Insight Journal, we did not add a Dart
> test since the available tests (ImageCompare and ImageCopy) did not seem
> to apply to our filter which inherits from the MeshToMeshFilter.  Are we
> correct in assuming that the ImageCompare test does not apply to our mesh
> output (i.e. will it also compare two meshes)?  Is there a MeshCompare
> test?
> 
> Thanks,
> 
> John Melonakos
> 
> _______________________________________________
> Insight-developers mailing list
> Insight-developers at itk.org
> http://www.itk.org/mailman/listinfo/insight-developers
> 
>