Feature mining for GMM-based speech quality measurement

Author

Falk, Tiago H. ; Chan, Wai-Yip

Author_Institution

Dept. of Electr. & Comput. Eng., Queen´´s Univ., Kingston, Ont., Canada

Volume

2

fYear

2004

fDate

7-10 Nov. 2004

Firstpage

2290

Abstract

We propose a novel approach to objective speech quality measurement using feature mining and Gaussian mixture models (GMMs). A large pool of perceptual distortion features is extracted from the speech signal and data mining techniques are used to sift out the most relevant feature variables from the pool. We examine using multivariate adaptive regression splines (MARS), classification and regression trees (CART), a hybrid CART-MARS scheme, and the sequential forward selection (SFS) algorithm for data mining. For our speech databases, the SFS algorithm provides best performance with a five-feature, three-component GMM. A reduction of 21.7% in root-mean-squared mean opinion score estimation error is obtained in comparison with ITU-T P.862 PESQ.

Keywords

Gaussian processes; audio databases; data mining; feature extraction; mean square error methods; regression analysis; signal classification; speech processing; splines (mathematics); trees (mathematics); GMM-based speech quality measurement; Gaussian mixture models; classification and regression trees; data mining techniques; feature mining; multivariate adaptive regression splines; perceptual distortion feature extraction; root-mean-squared mean opinion score estimation error; sequential forward selection; speech databases; speech signal; Classification tree analysis; Data mining; Distortion measurement; Electric variables measurement; Estimation error; Mars; Quality assessment; Regression tree analysis; Speech analysis; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Signals, Systems and Computers, 2004. Conference Record of the Thirty-Eighth Asilomar Conference on

Print_ISBN

0-7803-8622-1

Type

conf

DOI

10.1109/ACSSC.2004.1399576

Filename

1399576