• Jun 04, 2019 News!IJIEE Vol. 9, No. 2 issue has been published online!   [Click]
  • May 14, 2019 News!Vol.7, No.5-Vol.8, no.2 has been indexed by EI (Inspec).   [Click]
  • Apr 15, 2019 News!Papers published in Vol.8, No.1-No.4 have all received dois from Crossref.
General Information
    • ISSN: 2010-3719 (Online)
    • Abbreviated Title: Int. J. Inf. Electron. Eng.
    • Frequency: Quarterly
    • DOI: 10.18178/IJIEE
    • Editor-in-Chief: Prof. Chandratilak De Silva Liyanage
    • Executive Editor: Jennifer Zeng
    • Abstracting/ Indexing : Google Scholar, Electronic Journals Library, Crossref and ProQuest, Ei (INSPEC, IET), EBSCO.
    • E-mail ijiee@ejournal.net
Editor-in-chief

 
University of Brunei Darussalam, Brunei Darussalam   
" It is a great honor to serve as the editor-in-chief of IJIEE. I'll work together with the editorial team. Hopefully, The value of IJIEE will be well recognized among the readers in the related field."

IJIEE 2012 Vol.2(4): 552-555 ISSN: 2010-3719
DOI: 10.7763/IJIEE.2012.V2.159

A Combination of Maximum Likelihood Bayesian Framework and Discriminative Linear Transforms for Speaker Adaptation

Sh. Pirhosseinloo and Sh. Javadi
Abstract—Linear transforms are one of the most commonly used methods to speaker adaptation. In this paper, we present a combinational method of Bayesian framework and maximum likelihood linear regression as well as discriminative method for speaker adaptation. Furthermore significant gains can be obtained using discriminative training for acoustic models. Experiments on supervised adaptation on Persian data show that the combinational method outperforms both Maximum likelihood linear regression and Bayesian framework. Also the proposed method with discriminative adaptation outperforms previously proposed methods for transform estimation and discriminative training outperforms ML training.

Index Terms—Discriminative linear transforms, maximum-a-posterior adaptation, maximum likelihood linear regression adaptation, speech recognition, speaker adaptation.

Sh. Pirhosseinloo is the member of Scientific Association of Electrical and Electronic Engineering, Islamic Azad University Central Tehran Branch, and Tehran, Iran (e-mail: sh.pier@yahoo.com). 
Sh. Javadi is the assistant Professor of Electrical Engineering Department Islamic Azad University Central Tehran Branch, Tehran, Iran (e-mail: sh.javadi@iauctb.ac.ir).

[PDF]

Cite: Sh. Pirhosseinloo and Sh. Javadi, "A Combination of Maximum Likelihood Bayesian Framework and Discriminative Linear Transforms for Speaker Adaptation," International Journal of Information and Electronics Engineering vol. 2, no. 4, pp. 552-555, 2012.

Copyright © 2008-2019. International Journal of Information and Electronics Engineering. All rights reserved.
E-mail: ijiee@ejournal.net