A sequential feature selection algorithm for GMM-based speech quality estimation

Author

Falk, Tiago H. ; Wai-Yip Chan

Author_Institution

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

fYear

2005

fDate

4-8 Sept. 2005

Firstpage

1

Lastpage

4

Abstract

We propose a sequential feature selection algorithm for designing Gaussian mixture model (GMM) based estimators. Feature selection is performed progressively to minimize estimation errors. The algorithm is applied to design estimators of subjective speech quality. Simulation shows that estimators designed using the proposed algorithm outperform two benchmark algorithms by as much as 39% in correlation and 24% in root-mean-squared error. Furthermore, features selected by the proposed algorithm are suitable for diagonal GMM estimators, which incur lower computational complexity.

Keywords

Gaussian processes; computational complexity; mean square error methods; mixture models; speech processing; Gaussian mixture model; computational complexity; diagonal GMM estimators; estimation errors; root mean squared error; sequential feature selection; speech quality estimation; subjective speech quality; Algorithm design and analysis; Correlation; Covariance matrices; Estimation; Mars; Prediction algorithms; Speech;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing Conference, 2005 13th European

Conference_Location

Antalya

Print_ISBN

978-160-4238-21-1

Type

conf

Filename

7078016