Speaker identification using pairwise log-likelihood ratio measures

Author

Chao, Yi-Hsiang

Author_Institution

Dept. of Appl. Geomatics, Ching Yun Univ., Taoyuan, Taiwan

fYear

2012

fDate

29-31 May 2012

Firstpage

1248

Lastpage

1251

Abstract

T-norm and GMM-UBM are two predominant log-likelihood ratio (LLR)-based approaches for speaker verification in the last decade. In this paper, we embed T-norm and GMM-UBM in the new approaches based on cross likelihood ratio (CLR), named the pairwise LLR measures, for speaker identification tasks. This pairwise LLR measures can provide some extent of compensation for the conventional speaker identification method especially when client speaker models are not robust. Our experimental results show that the proposed pairwise LLR methods outperform the conventional GMM-UBM speaker identification approach.

Keywords

Gaussian processes; speaker recognition; Gaussian mixture model; T-norm; cross likelihood ratio; log-likelihood ratio-based approach; pairwise LLR measure; pairwise log-likelihood ratio measure; speaker identification; speaker verification; Adaptation models; Biological system modeling; Data models; Robustness; Speaker recognition; Speech; Vectors; Gaussian mixture model; cross likelihood ratio; log-likelihood ratio; speaker identification;

fLanguage

English

Publisher

ieee

Conference_Titel

Fuzzy Systems and Knowledge Discovery (FSKD), 2012 9th International Conference on

Conference_Location

Sichuan

Print_ISBN

978-1-4673-0025-4

Type

conf

DOI

10.1109/FSKD.2012.6234345

Filename

6234345