Rapid adaptation using linear spectral transformation for embedded speech recognisers

Y. Cho, Dongsuk Yook

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Embedded speech recognisers are typically used in unknown mobile environments where the acoustic conditions frequently change. Since a large amount of adaptation data is not usually available for such environments, the adaptation methods for the acoustic models of these recognisers must improve the recognition performance with only a small amount of adaptation data. In this Letter, we show that maximum likelihood linear spectral transformation provides the advantage of rapid adaptation using a very limited amount of adaptation data for the embedded acoustic models.

Original languageEnglish
Pages (from-to)1040-1042
Number of pages3
JournalElectronics Letters
Volume44
Issue number17
DOIs
Publication statusPublished - 2008 Aug 28

Fingerprint

Acoustics
Maximum likelihood

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Rapid adaptation using linear spectral transformation for embedded speech recognisers. / Cho, Y.; Yook, Dongsuk.

In: Electronics Letters, Vol. 44, No. 17, 28.08.2008, p. 1040-1042.

Research output: Contribution to journalArticle

@article{757a22aed79b436c837a5e94dec5c88b,
title = "Rapid adaptation using linear spectral transformation for embedded speech recognisers",
abstract = "Embedded speech recognisers are typically used in unknown mobile environments where the acoustic conditions frequently change. Since a large amount of adaptation data is not usually available for such environments, the adaptation methods for the acoustic models of these recognisers must improve the recognition performance with only a small amount of adaptation data. In this Letter, we show that maximum likelihood linear spectral transformation provides the advantage of rapid adaptation using a very limited amount of adaptation data for the embedded acoustic models.",
author = "Y. Cho and Dongsuk Yook",
year = "2008",
month = "8",
day = "28",
doi = "10.1049/el:20081503",
language = "English",
volume = "44",
pages = "1040--1042",
journal = "Electronics Letters",
issn = "0013-5194",
publisher = "Institution of Engineering and Technology",
number = "17",

}

TY - JOUR

T1 - Rapid adaptation using linear spectral transformation for embedded speech recognisers

AU - Cho, Y.

AU - Yook, Dongsuk

PY - 2008/8/28

Y1 - 2008/8/28

N2 - Embedded speech recognisers are typically used in unknown mobile environments where the acoustic conditions frequently change. Since a large amount of adaptation data is not usually available for such environments, the adaptation methods for the acoustic models of these recognisers must improve the recognition performance with only a small amount of adaptation data. In this Letter, we show that maximum likelihood linear spectral transformation provides the advantage of rapid adaptation using a very limited amount of adaptation data for the embedded acoustic models.

AB - Embedded speech recognisers are typically used in unknown mobile environments where the acoustic conditions frequently change. Since a large amount of adaptation data is not usually available for such environments, the adaptation methods for the acoustic models of these recognisers must improve the recognition performance with only a small amount of adaptation data. In this Letter, we show that maximum likelihood linear spectral transformation provides the advantage of rapid adaptation using a very limited amount of adaptation data for the embedded acoustic models.

UR - http://www.scopus.com/inward/record.url?scp=50049089781&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=50049089781&partnerID=8YFLogxK

U2 - 10.1049/el:20081503

DO - 10.1049/el:20081503

M3 - Article

VL - 44

SP - 1040

EP - 1042

JO - Electronics Letters

JF - Electronics Letters

SN - 0013-5194

IS - 17

ER -