Tree-based Gaussian mixture models for speaker verification
(2005-12)
Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2005.
Thesis
The Gaussian mixture model (GMM) performs very effectively in applications such as speech and speaker recognition. However, evaluation speed is greatly reduced when the GMM has a large number of mixture components. Various techniques improve the evaluation speed by reducing the number of required Gaussian evaluations.