GMM-based matching ability measurement of a speech recognizer and a feature set

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

In this work, we propose a Gaussian mixture model-based recognizer selection method to overcome the acoustic mismatch between training and testing environments of a speech recognition system. The method evaluates the preference of a system over other for a specific feature set. By applying it to compare the two speech recognition systems constructed with wireline speech and wireless speech, respectively, it is shown that the matched condition of wireless training and testing can give better recognition accuracies than the mismatched condition.

Original languageEnglish
Title of host publicationFuture Communication, Computing, Control and Management
Pages377-383
Number of pages7
EditionVOL. 2
DOIs
StatePublished - 2012
Event2011 International Conference on Future Communication, Computing, Control and Management, ICF4C 2011 - Phuket, Thailand
Duration: 16 Dec 201117 Dec 2011

Publication series

NameLecture Notes in Electrical Engineering
NumberVOL. 2
Volume142 LNEE
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

Conference2011 International Conference on Future Communication, Computing, Control and Management, ICF4C 2011
Country/TerritoryThailand
CityPhuket
Period16/12/1117/12/11

Keywords

  • acoustic mismatch
  • Gaussian mixture model
  • recognizer selection

Fingerprint

Dive into the research topics of 'GMM-based matching ability measurement of a speech recognizer and a feature set'. Together they form a unique fingerprint.

Cite this