Feature transform in linear spectral domain for fast channel adaptation

D. Kim, Dongsuk Yook

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

A feature transform algorithm is proposed using the maximum likelihood spectral transform method for robust speech recognition. The feature transform compensates for the mismatch between clean training data and noisy testing data. A new technique is proposed to estimate a transform matrix for the feature transform in the linear spectral domain with a small number of parameters, which allows fast nonlinear adaptation to be realised.

Original languageEnglish
Pages (from-to)1313-1315
Number of pages3
JournalElectronics Letters
Volume40
Issue number20
DOIs
Publication statusPublished - 2004 Sep 30

Fingerprint

Speech recognition
Maximum likelihood
Testing

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Feature transform in linear spectral domain for fast channel adaptation. / Kim, D.; Yook, Dongsuk.

In: Electronics Letters, Vol. 40, No. 20, 30.09.2004, p. 1313-1315.

Research output: Contribution to journalArticle

@article{74ca4fc514ce48f48683d0b1db4dcf57,
title = "Feature transform in linear spectral domain for fast channel adaptation",
abstract = "A feature transform algorithm is proposed using the maximum likelihood spectral transform method for robust speech recognition. The feature transform compensates for the mismatch between clean training data and noisy testing data. A new technique is proposed to estimate a transform matrix for the feature transform in the linear spectral domain with a small number of parameters, which allows fast nonlinear adaptation to be realised.",
author = "D. Kim and Dongsuk Yook",
year = "2004",
month = "9",
day = "30",
doi = "10.1049/el:20045484",
language = "English",
volume = "40",
pages = "1313--1315",
journal = "Electronics Letters",
issn = "0013-5194",
publisher = "Institution of Engineering and Technology",
number = "20",

}

TY - JOUR

T1 - Feature transform in linear spectral domain for fast channel adaptation

AU - Kim, D.

AU - Yook, Dongsuk

PY - 2004/9/30

Y1 - 2004/9/30

N2 - A feature transform algorithm is proposed using the maximum likelihood spectral transform method for robust speech recognition. The feature transform compensates for the mismatch between clean training data and noisy testing data. A new technique is proposed to estimate a transform matrix for the feature transform in the linear spectral domain with a small number of parameters, which allows fast nonlinear adaptation to be realised.

AB - A feature transform algorithm is proposed using the maximum likelihood spectral transform method for robust speech recognition. The feature transform compensates for the mismatch between clean training data and noisy testing data. A new technique is proposed to estimate a transform matrix for the feature transform in the linear spectral domain with a small number of parameters, which allows fast nonlinear adaptation to be realised.

UR - http://www.scopus.com/inward/record.url?scp=6444225123&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=6444225123&partnerID=8YFLogxK

U2 - 10.1049/el:20045484

DO - 10.1049/el:20045484

M3 - Article

VL - 40

SP - 1313

EP - 1315

JO - Electronics Letters

JF - Electronics Letters

SN - 0013-5194

IS - 20

ER -