Publication Details

Language Recognition in iVectors Space

MARTÍNEZ GONZÁLEZ, D.; PLCHOT, O.; BURGET, L.; GLEMBEK, O.; MATĚJKA, P. Language Recognition in iVectors Space. In Proceedings of Interspeech 2011. Proceedings of Interspeech. Florence: International Speech Communication Association, 2011. p. 861-864. ISBN: 978-1-61839-270-1. ISSN: 1990-9772.
Czech title
Rozpoznávání jazyka v prostoru iVektorů
Type
conference paper
Language
English
Authors
Martínez González David
Plchot Oldřich, Ing., Ph.D. (DCGM)
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
Glembek Ondřej, Ing., Ph.D.
Matějka Pavel, Ing., Ph.D. (DCGM)
URL
Keywords

Acoustic Language Recognition, iVectors, Joint Factor Analysis

Abstract

We have introduced a novel approach for language recognition.Three classifiers (linear generative model, SVM and logistic regression)have been tested in the iVector space, and all outperformthe state-of-the-art JFA system. Very simple and fast classifierbased on linear generative model provides excellent performanceover all conditions. The advantage of this classifieris also its scalability: addition of a new language only requiresestimating the mean over the corresponding iVectors. Most ofthe computational load is in the iVector generation. Hence, as anext step, we will try to obtain iVectors from the utterances andthe corresponding sufficient statistics in a more direct way.

Annotation

The concept of so called iVectors, where each utterance is represented by fixed-length low-dimensional feature vector, has recently become very successfully in speaker verification. In this work, we apply the same idea in the context of Language Recognition (LR). To recognize language in the iVector space, we experiment with three different linear classifiers: one based on a generative model, where classes are modeled by Gaussian distributions with shared covariance matrix, and two discriminative classifiers, namely linear Support Vector Machine and Logistic Regression. The tests were performed on the NIST LRE 2009 dataset and the results were compared with stateof- the-art LR based on Joint Factor Analysis (JFA). While the iVector system offers better performance, it also seems to be complementary to JFA, as their fusion shows another improvement.

Published
2011
Pages
861–864
Journal
Proceedings of Interspeech, vol. 2011, no. 8, ISSN 1990-9772
Proceedings
Proceedings of Interspeech 2011
Conference
Interspeech Conference, Florence, IT
ISBN
978-1-61839-270-1
Publisher
International Speech Communication Association
Place
Florence
EID Scopus
BibTeX
@inproceedings{BUT76437,
  author="David {Martínez González} and Oldřich {Plchot} and Lukáš {Burget} and Ondřej {Glembek} and Pavel {Matějka}",
  title="Language Recognition in iVectors Space",
  booktitle="Proceedings of Interspeech 2011",
  year="2011",
  journal="Proceedings of Interspeech",
  volume="2011",
  number="8",
  pages="861--864",
  publisher="International Speech Communication Association",
  address="Florence",
  isbn="978-1-61839-270-1",
  issn="1990-9772",
  url="http://www.fit.vutbr.cz/research/groups/speech/publi/2011/martinez_interspeech2011_291.pdf"
}
Back to top