is k-means++ essential a speed up for k-means clustering? is it already integrated in vtkKMeansStatistics? is there anything in vtk (or otherwise can anyone recommend) something that will easily find a suitable number of clusters. Liam