[Insight-users] Morphological Filter boundary condition and kernel size

[Karthik Krishnan]

Hi Quan,

Did you try it on the images in 
Testing/Data/Baseline/Baseline/BasicFilters/itkGrayscaleFunction*.mha. 
The filter in the testing repository did not seem to change the origin 
of the image. Try adding a line such as "Offset = 0 0" to one of the mha 
images to check. Maybe I am wrong.

Did you imply that you would rather have the image change in size, due 
to dilation operation. Currently due to the way boundary conditions are 
defined on the neighborhood iterators, the image doesn't really change 
in size and is assumed to have 0 outside the bounds. 
Although I am still confused cause opening/closing closing operations 
shouldn't change the size of the image.

Thanks
Karthik

Quan Chen wrote:

> First, for all those Morphological or Geodesic filters, the structure 
> element is in the unit of pixel only?  I am wondering if it is a good 
> idea to pass the image information (pixel size) to StructureElement so 
> that it will accept the size in actual dimensions like other filter 
> (like Gaussian) did. 
>  
> Secondly, I tried to use the GrayscaleMorphologicalOpeningImageFilter 
> that itk add recently.  I noticed that the filtered image shifted 2*n 
> pixels in all dimensions when I use the structure element with a 
> radius of n.  I am wondering if it is the filter design bug or I 
> didn't compile the updated version correctly (mixed with old version?).
>  
>
>------------------------------------------------------------------------
>
>_______________________________________________
>Insight-users mailing list
>Insight-users at itk.org
>http://www.itk.org/mailman/listinfo/insight-users
>  
>