Publication detail

A Region-specific Feature-space Transformation for Speaker Adaptation and Singularity Analysis of Jacobian Matrix

RATH, S. BURGET, L. KARAFIÁT, M. GLEMBEK, O. ČERNOCKÝ, J.

Original Title

A Region-specific Feature-space Transformation for Speaker Adaptation and Singularity Analysis of Jacobian Matrix

Type

article in a collection out of WoS and Scopus

Language

English

Original Abstract

This paper describes the difficulties associated with soft R-FMLLR. By analyzing the Jacobian matrix, it was concluded that the transformation is most likely to be non-invertible and in this case ML estimation adversely affects the performance. A new transformation, hard R-FMLLR, is presented. It is shown that the performance of the proposed method is better than soft R-FMLLR and it is computationally more efficient.

Keywords

speaker recognition, speaker adaptation, feature-space transformation, speech recognition

Authors

RATH, S.; BURGET, L.; KARAFIÁT, M.; GLEMBEK, O.; ČERNOCKÝ, J.

RIV year

2013

Released

25. 8. 2013

Publisher

International Speech Communication Association

Location

Lyon

ISBN

978-1-62993-443-3

Book

Proceedings of Interspeeech 2013

ISBN

2308-457X

Periodical

Proceedings of the 14th Annual Conference of the International Speech Communication Association (Interspeech 2013).

Number

8

State

French Republic

Pages from

1228

Pages to

1232

Pages count

5

URL

http://www.fit.vutbr.cz/research/groups/speech/publi/2013/rath_interspeech2013_IS130146.pdf

BibTex

@inproceedings{BUT103551,
  author="Shakti Prasad {Rath} and Lukáš {Burget} and Martin {Karafiát} and Ondřej {Glembek} and Jan {Černocký}",
  title="A Region-specific Feature-space Transformation for Speaker Adaptation and Singularity Analysis of Jacobian Matrix",
  booktitle="Proceedings of Interspeeech 2013",
  year="2013",
  journal="Proceedings of the 14th Annual Conference of the International Speech Communication Association (Interspeech 2013).",
  number="8",
  pages="1228--1232",
  publisher="International Speech Communication Association",
  address="Lyon",
  isbn="978-1-62993-443-3",
  issn="2308-457X",
  url="http://www.fit.vutbr.cz/research/groups/speech/publi/2013/rath_interspeech2013_IS130146.pdf"
}