[Insight-developers] VectorConnectedComponent filter?

[David Doria]

On Wed, Dec 28, 2011 at 8:33 AM, Bradley Lowekamp <blowekamp at mail.nih.gov>wrote:

> Hello,
>
> I am have wrapped the VectorConnectdComponentImageFilter for SimpleITK.
> However, I am unable to figure how this filter would be useful and what a
> good example test would be for it. Here is the doxygen page:
>
>
> http://www.itk.org/Doxygen/html/classitk_1_1VectorConnectedComponentImageFilter.html
>
> This filter expect the input vector's of the image to be normalize. Then
> it compare neighboring pixel by computing 1-v1 dot v2. I am just wondering
> what the use case is for this filter. Perhaps some diffusion images? And if
> some one has a good test image for input to this filter?
>

I'm sure this was not the intended use (since it is non-medical) but this
filter could be used for surface segmentation of a depth image. I could
make an image of the normals of a view (3 sides) of a cube to use for
testing if you don't come up with a different usage.


> Related to this filter requiring normalized vector's in the input image.
> We don't have a filter to normalize the vectors in an image. This should be
> simple enough to do with a functor based filter. Just wondering if other
> people think this is an needed missing filter, because it makes generating
> a synthetic input image for this filter rather tedious.
>

Yes, I would use this addition. I have a NormalizeVectorImage function in
my Helpers namespace that I use all the time.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/mailman/private/insight-developers/attachments/20111228/e13bac26/attachment.htm>