An unsupervised approach to language identification

Author

Pellegrino, F. ; André-Obrecht, R.

Author_Institution

IRIT, Toulouse, France

Volume

2

fYear

1999

fDate

15-19 Mar 1999

Firstpage

833

Abstract

This paper presents an unsupervised approach to automatic language identification (ALI) based on vowel system modeling. Each language vowel system is modeled by a Gaussian mixture model (GMM) trained with automatically detected vowels. Since this detection is unsupervised and language independent, no labeled data are required. GMMs are initialized using an efficient data-driven variant of the LBG algorithm: the LBG-Rissanen (1983) algorithm. With 5 languages from the OGI MLTS corpus and in a close set identification task, we reach 79% of correct identification using only the vowel segments detected in 45 second duration utterances for the male speakers

Keywords

Gaussian processes; acoustic signal processing; natural languages; speech processing; unsupervised learning; 45 s; Gaussian mixture model; LBG-Rissanen algorithm; OGI MLTS corpus; acoustic processing; automatic speech processing; automatically detected vowels; close set identification task; correct identification; data-driven variant LBG algorithm; language identification; language independent detection; language vowel system; male speakers; unsupervised detection; utterances; vowel segments; vowel system modeling; Acoustic signal detection; Cepstral analysis; Databases; Entropy; Hidden Markov models; Modeling; Natural languages; Speech processing; Speech recognition; Topology;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1999. Proceedings., 1999 IEEE International Conference on

Conference_Location

Phoenix, AZ

ISSN

1520-6149

Print_ISBN

0-7803-5041-3

Type

conf

DOI

10.1109/ICASSP.1999.759800

Filename

759800