An effective identification method for speaker recognition based on PCA and double VQ

Recognition rate is always the research focus in speaker recognition, however, recognition response speed also is the key which determines the practices of the system. For the phenomenon that recognition response speed is slow in large crowds, in this paper, a new recognition method is proposed, whi...

Full description

Saved in:
Bibliographic Details
Published in2009 International Conference on Machine Learning and Cybernetics Vol. 3; pp. 1686 - 1689
Main Authors Zhen-Dong Zhao, Jing Zhang, Jing-Feng Tian, Yun-Yong Lou
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.07.2009
Subjects
Covariance matrix
Cybernetics
Delay
Double Vector Quantization
Eigenvalues and eigenfunctions
Machine learning
Principal component analysis
Real time systems
Space technology
Speaker recognition
Vector quantization
Online AccessGet full text
ISBN9781424437023
1424437024
ISSN2160-133X
DOI10.1109/ICMLC.2009.5212249

Cover

More Information
Summary:Recognition rate is always the research focus in speaker recognition, however, recognition response speed also is the key which determines the practices of the system. For the phenomenon that recognition response speed is slow in large crowds, in this paper, a new recognition method is proposed, which takes advantage of that PCA classifier is simple and double VQ is real-time response: Firstly, PCA Classifier is used for preliminary classification, then make the final distinguishing using double VQ. In this way, the system' recognition time greatly reduced. With the growth of registration speakers, compared with conventional speaker recognition method, the advantage of this method will be more apparent.
ISBN:9781424437023
1424437024
ISSN:2160-133X
DOI:10.1109/ICMLC.2009.5212249