Fuzzy expectation-maximisation algorithm for speech and speaker recognition

Author

Tran, Dat ; Wagner, Michael

Author_Institution

Sch. of Comput., Canberra Univ., Belconnen, ACT, Australia

fYear

1999

fDate

36342

Firstpage

421

Lastpage

425

Abstract

A fuzzy c-means approach to the expectation-maximisation (EM) algorithm is proposed. A family of fuzzy EM algorithms of various degrees of fuzziness is presented, where the EM algorithm is referred to as a fuzzy EM algorithm with degree of fuzziness of one. This fuzzy approach can be applied to EM-style algorithms such as the Baum-Welch algorithm for hidden Markov models, and the EM algorithm for Gaussian mixture models in speech and speaker recognition. The fuzzy EM algorithm for Gaussian mixture models is considered in detail as a demonstration for applying the fuzzy EM algorithm

Keywords

Gaussian processes; fuzzy set theory; hidden Markov models; maximum likelihood estimation; optimisation; speaker recognition; Baum-Welch algorithm; EM-style algorithms; Gaussian mixture models; fuzzy EM algorithms; fuzzy c-means approach; fuzzy expectation-maximisation algorithm; hidden Markov models; speaker recognition; speech recognition; Biological system modeling; Clustering algorithms; Expectation-maximization algorithms; Extraterrestrial measurements; Hidden Markov models; Inference algorithms; Iterative algorithms; Maximum likelihood estimation; Speaker recognition; Speech;

fLanguage

English

Publisher

ieee

Conference_Titel

Fuzzy Information Processing Society, 1999. NAFIPS. 18th International Conference of the North American

Conference_Location

New York, NY

Print_ISBN

0-7803-5211-4

Type

conf

DOI

10.1109/NAFIPS.1999.781727

Filename

781727