Optimization of Gaussian mixture model parameters for speaker identification

Q. Y. HONG*, Sam KWONG, H. L. WANG

*Corresponding author for this work

Research output: Book Chapters | Papers in Conference ProceedingsConference paper (refereed)Researchpeer-review

4 Citations (Scopus)

Abstract

Gaussian mixture model (GMM) [1] has been widely used for modeling speakers. In speaker identification, one major problem is how to generate a set of GMMs for identification purposes based upon the training data. Due to the hill-climbing characteristic of the maximum likelihood (ML) method, any arbitrary estimate of the initial model parameters will usually lead to a sub-optimal model in practice. To resolve this problem, this paper proposes a hybrid training method based on the genetic algorithm (GA). It utilizes the global searching capability of the GA and combines the effectiveness of the ML method.

Original languageEnglish
Title of host publicationGenetic and Evolutionary Computation : GECCO 2004 : Genetic and Evolutionary Computation Conference, Seattle, WA, USA, June 26–30, 2004 Proceedings, Part II
EditorsRiccardo POLI, Owen HOLLAND, Wolfgang BANZHAF, Hans-Georg BEYER, Edmund BURKE, Paul DARWEN, Dipankar DASGUPTA, Dario FLOREANO, James FOSTER, Mark HARMAN, Pier Luca LANZI, Lee SPECTOR, Andrea G. B. TETTAMANZI, Dirk THIERENS, Andrew M. TYRRELL
Place of PublicationHeidelberg
PublisherSpringer Berlin
Pages1310-1311
Number of pages2
ISBN (Electronic)9783540248552
ISBN (Print)9783540223436
DOIs
Publication statusPublished - 2004
Externally publishedYes
Event2004 Genetic and Evolutionary Computation Conference - Seattle, United States
Duration: 26 Jun 200430 Jun 2004

Publication series

NameLecture Notes in Computer Science
Volume3103
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2004 Genetic and Evolutionary Computation Conference
Abbreviated titleGECCO 2004
Country/TerritoryUnited States
CitySeattle
Period26/06/0430/06/04

Keywords

  • Genetic Algorithm
  • Gaussian Mixture Model
  • Speaker Identification
  • Female Speaker
  • Speaker Model

Fingerprint

Dive into the research topics of 'Optimization of Gaussian mixture model parameters for speaker identification'. Together they form a unique fingerprint.

Cite this