A set of points are lying in a plane given by its normal vector and a point in space. I can see that those points are regrouped together in a unpredictable way: some are alone some groups counts more.
My question is what algorithm should I use to cluster such a set of point. I've already looked into k-mean class of algorithm but I can't know the number of clusters (k) by advance.
One obvious solution would be to brutally compute all distances between all points but it seems inefficient with the data structure I have in mind. Maybe using a tree representation ? If so I would appreciate any documentation or help on the subject. I am using C++11.
Aucun commentaire:
Enregistrer un commentaire