Publication Details

Bilinear Factor Analysis for iVector Based Speaker Verification

LEI Yun, BURGET Lukáš and SCHEFFER Nicolas. Bilinear Factor Analysis for iVector Based Speaker Verification. In: Proceedings of Interspeech. Portland, Oregon: International Speech Communication Association, 2012, pp. 1-4. ISBN 978-1-62276-759-5.
Czech title
Bililneární faktorová analýza pro ověřování mluvčího založené na iVektorech
Type
conference paper
Language
english
Authors
Lei Yun (SRI)
Burget Lukáš, doc. Ing., Ph.D. (DCGM FIT BUT)
Scheffer Nicolas (SRI)
URL
Keywords

speaker recognition, i-Vectors, PLDA

Abstract

In this study, we have proposed and tested a new extension of the PLDA model, where within-class (channel) variability is modeled as a function of the class (speaker) location in the feature (iVector) space.

Annotation

The combination of iVector extraction and Probabilistic Linear Discriminant Analysis (PLDA) model forms a basis of the current state of the art speaker verification. The PLDA model makes an assumption that the within-speaker (or inter-session) variability in the iVector space is independent of speaker identity. In this work we propose a new model, which can be seen as an extension of PLDA, relaxing this assumption and allowing the within-speaker variability to be different for different locations of speakers in the iVector space. The potential of the proposed model is demonstrated in preliminary experiments.

Published
2012
Pages
1-4
Proceedings
Proceedings of Interspeech
Conference
Interspeech Conference, Portland, US
ISBN
978-1-62276-759-5
Publisher
International Speech Communication Association
Place
Portland, Oregon, US
BibTeX 
@INPROCEEDINGS{FITPUB10778,
   author = "Yun Lei and Luk\'{a}\v{s} Burget and Nicolas Scheffer",
   title = "Bilinear Factor Analysis for iVector Based Speaker Verification",
   pages = "1--4",
   booktitle = "Proceedings of Interspeech",
   year = 2012,
   location = "Portland, Oregon, US",
   publisher = "International Speech Communication Association",
   ISBN = "978-1-62276-759-5",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/10778"
}
Back to top