Publication Details

Deep Neural Networks and Hidden Markov Models in i-vector-based Text-Dependent Speaker Verification

ZEINALI, H.; BURGET, L.; SAMETI, H.; GLEMBEK, O.; PLCHOT, O. Deep Neural Networks and Hidden Markov Models in i-vector-based Text-Dependent Speaker Verification. In Proceedings of Odyssey 2016, The Speaker and Language Recognition Workshop. Proceedings of Odyssey: The Speaker and Language Recognition Workshop Odyssey 2014, Joensuu, Finland. Bilbao: International Speech Communication Association, 2016. p. 24-30. ISSN: 2312-2846.
Czech title
Hluboké neuronové sítě a skryté Markovovy modely v i-vektorovém systému pro ověřování mluvčího závislém na textu
Type
conference paper
Language
English
Authors
Zeinali Hossein, Ph.D. (DCGM)
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
Sameti Hossein (FIT)
Glembek Ondřej, Ing., Ph.D.
Plchot Oldřich, Ing., Ph.D. (DCGM)
URL
Keywords

deep neural networks,  hidden Markov Models, i-vector-based, text-dependent, speaker verification

Abstract

This article is about deep neural networks and hidden Markov models in i-vector-based text-dependent speaker verification.

Annotation

Techniques making use of Deep Neural Networks (DNN) have recently been seen to bring large improvements in textindependent speaker recognition. In this paper, we verify that the DNN based methods result in excellent performances in the context of text-dependent speaker verification as well. We build our system on the previously introduced HMM based ivector approach, where phone models are used to obtain frame level alignment in order to collect sufficient statistics for ivector extraction. For comparison, we experiment with an alternative alignment obtained directly from the output of DNN trained for phone classification. We also experiment with DNN based bottleneck features and their combinations with standard cepstral features. Although the i-vector approach is generally considered not suitable for text-dependent speaker verification, we show that our HMM based approach combined with bottleneck features provides truly state-of-the-art performance on RSR2015 data.

Published
2016
Pages
24–30
Journal
Proceedings of Odyssey: The Speaker and Language Recognition Workshop Odyssey 2014, Joensuu, Finland, vol. 2016, no. 06, ISSN 2312-2846
Proceedings
Proceedings of Odyssey 2016, The Speaker and Language Recognition Workshop
Publisher
International Speech Communication Association
Place
Bilbao
DOI
10.21437/Odyssey.2016-4
EID Scopus
2-s2.0-85048307659
BibTeX 
@inproceedings{BUT131003,
  author="Hossein {Zeinali} and Lukáš {Burget} and Hossein {Sameti} and Ondřej {Glembek} and Oldřich {Plchot}",
  title="Deep Neural Networks and Hidden Markov Models in i-vector-based Text-Dependent Speaker Verification",
  booktitle="Proceedings of Odyssey 2016, The Speaker and Language Recognition Workshop",
  year="2016",
  journal="Proceedings of Odyssey: The Speaker and Language Recognition Workshop Odyssey 2014, Joensuu, Finland",
  volume="2016",
  number="06",
  pages="24--30",
  publisher="International Speech Communication Association",
  address="Bilbao",
  doi="10.21437/Odyssey.2016-4",
  issn="2312-2846",
  url="http://www.odyssey2016.org/papers/pdfs_stamped/63.pdf"
}
Files
Back to top