[Insight-developers] Bug in KdTreeBasedKmeansEstimator ?

[David Doria]

I didn't want to file a bug report yet because I'm not 100% sure I'm
using it right. In this example:

http://www.itk.org/Wiki/ITK/Examples/Statistics/KdTreeBasedKmeansEstimator

You'll see I am trying to cluster 3D points. I believe this line:

EstimatorType::ParametersType initialMeans(2*3);

means that I am trying to produce 2 clusters? Assuming that is
correct, the question is that if I add 10 points to the sample (which
I have hard coded into the example) it runs and gives a reasonable
looking solution. However, if I add 20 points (again hard coded, but
commented), I get this error:

  what():  /home/doriad/src/ITK/Modules/Numerics/Statistics/include/itkEuclideanDistanceMetric.hxx:66:
itk::ERROR: EuclideanDistanceMetric(0x8156578): The two measurement
vectors have unequal size (3 and 0)

Any thoughts as to what would cause this?

Thanks,

David