Mean shift algorithm for exponential families with applications to speaker clustering

Themos Stafylakis, Vassilis Katsouros, Patrick Kenny, Pierre Dumouchel

This work extends the mean shift algorithm from the observation space to the manifolds of parametric models that are formed by exponential families. We show how the Kullback-Leibler divergence and its dual define the corresponding affine connection and propose a method for incorporating the uncertainty in estimating the parameters. Experiments are carried out for the problem of speaker clustering, using both single Gaussians and i-vectors.

Cite as: Stafylakis, T., Katsouros, V., Kenny, P., Dumouchel, P. (2012) Mean shift algorithm for exponential families with applications to speaker clustering. Proc. The Speaker and Language Recognition Workshop (Odyssey 2012), 324-329

@inproceedings{stafylakis12b_odyssey,
  author={Themos Stafylakis and Vassilis Katsouros and Patrick Kenny and Pierre Dumouchel},
  title={{Mean shift algorithm for exponential families with applications to speaker clustering}},
  year=2012,
  booktitle={Proc. The Speaker and Language Recognition Workshop (Odyssey 2012)},
  pages={324--329}
}