[Insight-users] How to use KmeansEstimator class

[Luis Ibanez]

Hi Liu,

Please read the ITK Software Guide

  http://www.itk.org/ItkSoftwareGuide.pdf

in particular the Statistics Chapter.


---


You will find that the K-Means algorithm is
implemented independently of the data container.
The actual algorithm is executed using a List of
Samples as input.

The K-Means filter that operates over images,
uses an Image to List adaptor that presents
the pixels as a list of measurement vector.

http://www.itk.org/Insight/Doxygen/html/classitk_1_1Statistics_1_1ImageToListAdaptor.html



If you want to do K-Means on your special data,
you should write an equivalent adaptor. Let's
say a:

           SpecialImageToListAdaptor


Once you have this class, you can take advantage
of the K-Means algorithm built into the Statistics
framework.

For this last step you may want to follow the model
of the filters:

http://www.itk.org/Insight/Doxygen/html/classitk_1_1ScalarImageKmeansImageFilter.html



If you write such a class, you should consider
posting it as a Paper to the Insight Journal:

http://www.insightsoftwareconsortium.org/InsightJournal/

so it can recieve reviews from the community, and
eventually be included in the next release of ITK.



   Regards,


      Luis



--------------------
liu jianfei wrote:
> Hi,
> 
>    Current, I want to implement K-means clustering algorithm on my
> special data. The input data are all boundary voxels of a 3d volmetric
> dataset. They are connected according to 6-/18-/26- connected
> relationship.  The K-means clustering algorithm is performened in this
> dataset. I've check the some KmeansEstimator classes in ITK.  However,
> their input dataset is required to be a regular image. Is it possilbe
> for me to use ITK k-means clustering class? Thanks!
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
> 
>