kmeans() only accepts 2D and CV_32F (Feature #2307)

Added by Hilton Bristow over 4 years ago. Updated over 1 year ago.

Status:Open Start date:2012-08-26
Priority:Normal Due date:
Assignee:- % Done:


Target version:Next Hackathon
Difficulty: Pull request:


I would like to request that kmeans() within core/clustering be extended to support all cv datatypes (or at least double precision floating point).

It would also be good to extend the implementation to arbitrary dimensional data as well, and remove the dependency on using Vec2f types. Since the current implementation does not use a kdtree for storage, this should require only minimal changes.


Updated by Stefano Fabri over 4 years ago


Updated by Vadim Pisarevsky over 4 years ago

  • Target version deleted ()
  • Assignee deleted (Vadim Pisarevsky)

Updated by Marina Kolpakova over 4 years ago

  • Tracker changed from Bug to Feature

Updated by Kirill Kornyakov over 4 years ago

  • Target version set to Next Hackathon

Updated by Sebastian Krämer over 4 years ago

+1 for double support.

About the input format: isn't it common to require one feature vector per row? How should kmeans cope with data of arbitrary dimensionality?

Updated by Maksim Shabunin over 1 year ago

Issue has been transferred to GitHub:

Also available in: Atom PDF