Home // ICDT 2011, The Sixth International Conference on Digital Telecommunications // View article

Language Recognition With Locality Preserving Projection

Authors:
Jinchao Yang

Keywords: language recognition, language total variability, PCA, LDA, LPP, SVM

Abstract:
In this paper, we introduce locality preserving projection (LPP) to language recognition under the support vector machine (SVM) framework. The success of the use of total variability in language recognition shows that the global structure and linear manifold preserve discriminative language dependent information. The proposed LPP language recognition system believes the local structure and nonlinear manifold also contain discriminative language dependent information. Experiment results on 2007 National Institute of Standards and Technology (NIST) language Recognition Evaluation (LRE) databases show LPP language recognition system combining total variability language recognition system gains relative improvement in EER of 11.7% and in minDCF of 9.6% comparing to total variability language recognition system in 30-second tasks, and further improvement is obtained combining with state-of-the-art systems. It leads to gains of 13.8% in EER and 20.2% in minDCF compare with the performance of the combination of the MMI and the GMM-SVM systems.

Pages: 51 to 55

Copyright: Copyright (c) IARIA, 2011

Publication date: April 17, 2011

Published in: conference

ISSN: 2308-3964

ISBN: 978-1-61208-127-4

Location: Budapest, Hungary

Dates: from April 17, 2011 to April 22, 2011