Publication Details

Domain adaptation via within-class covariance correction in I-vector based speaker recognition systems

GLEMBEK, O.; MA, J.; MATĚJKA, P.; ZHANG, B.; PLCHOT, O.; BURGET, L.; MATSOUKAS, S. Domain adaptation via within-class covariance correction in I-vector based speaker recognition systems. In Proceedings of ICASSP 2014. Florencie: IEEE Signal Processing Society, 2014. p. 4060-4064. ISBN: 978-1-4799-2892-7.
Czech title
Adaptace na doménu pomocí vnitro-třídní kovarianční opravy v systému pro rozpoznávání mluvčího založeném na i-vektorech
Type
conference paper
Language
English
Authors
Glembek Ondřej, Ing., Ph.D.
Ma Jeff
Matějka Pavel, Ing., Ph.D.
Zhang Bing
Plchot Oldřich, Ing., Ph.D. (DCGM)
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
Matsoukas Spyros
URL
Keywords

speaker recognition, i-vectors, source normalization,LDA, inter-dataset variability compensation

Abstract

In this paper, we have shown a technique of within-class correctionfor Linear Discriminant Analysis estimation. We have shown thatwhen correct dataset clustering is used, adapting the within-classcovariance of LDA by low-rank between-dataset covariance matrixcan lead to significant improvement of the system, namely up to70% in the Domain Adaptation Task, and 17.5% and 36% relativein the RATS unmatched and semi-matched tasks, respectively. Thedataset clustering problem gave us an interesting direction for futureresearch.

Annotation

In this paper we propose a technique of Within-Class Covariance Correction (WCC) for Linear Discriminant Analysis (LDA) in Speaker Recognition to perform an unsupervised adaptation of LDA to an unseen data domain, and/or to compensate for speaker population difference among different portions of LDA training dataset. The paper follows on the study of source-normalization and interdatabase variability compensation techniques which deal with multimodal distribution of i-vectors. On the DARPA RATS (Robust Automatic Transcription of Speech) task, we show that, with two hours of unsupervised data, we improve the Equal-Error Rate (EER) by 17.5%, and 36% relative on the unmatched and semi-matched conditions, respectively. On the Domain Adaptation Challenge we show up to 70% relative EER reduction and we propose a data clustering procedure to identify the directions of the domain-based variability in the adaptation data.

Published
2014
Pages
4060–4064
Proceedings
Proceedings of ICASSP 2014
Conference
The 39th International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Florencie, IT
ISBN
978-1-4799-2892-7
Publisher
IEEE Signal Processing Society
Place
Florencie
DOI
10.1109/ICASSP.2014.6854359
UT WoS
000343655304011
EID Scopus
2-s2.0-84905268753
BibTeX 
@inproceedings{BUT111543,
  author="Ondřej {Glembek} and Jeff {Ma} and Pavel {Matějka} and Bing {Zhang} and Oldřich {Plchot} and Lukáš {Burget} and Spyros {Matsoukas}",
  title="Domain adaptation via within-class covariance correction in I-vector based speaker recognition systems",
  booktitle="Proceedings of ICASSP 2014",
  year="2014",
  pages="4060--4064",
  publisher="IEEE Signal Processing Society",
  address="Florencie",
  doi="10.1109/ICASSP.2014.6854359",
  isbn="978-1-4799-2892-7",
  url="https://www.fit.vut.cz/research/publication/10555/"
}
Files
Back to top