HMM Parameter Adaptation Using the Truncated First-Order VTS and EM Algorithm for Robust Speech Recognition

This paper presents a framework of HMM parameter adaptation technique for improving automatic speech recognition (ASR) performance in the noisy environments, which online combines the clean hidden Markov models (HMMs) with the noise model. Based on the given composite HMM corresponding to the initia...

Full description

Saved in:
Bibliographic Details
Published inComputational Intelligence and Security pp. 979 - 984
Main Authors Shen, Haifeng, Li, Qunxia, Guo, Jun, Liu, Gang
Format Book Chapter
LanguageEnglish
Published Berlin, Heidelberg Springer Berlin Heidelberg 2005
SeriesLecture Notes in Computer Science
Subjects
Automatic Speech Recognition
Noise Model
Noisy Environment
Recognition Rate
Speech Recognition
Online AccessGet full text
ISBN9783540308188
3540308180
ISSN0302-9743
1611-3349
DOI10.1007/11596448_145

Cover

More Information
Summary:This paper presents a framework of HMM parameter adaptation technique for improving automatic speech recognition (ASR) performance in the noisy environments, which online combines the clean hidden Markov models (HMMs) with the noise model. Based on the given composite HMM corresponding to the initial recognition pass result and truncated vector Taylor series, the noise model in the cepstral domain is updated and refined using iterative Expectation-Maximization (EM) algorithm under maximum likelihood (ML) criterion. Experiments results show that the presented approach in this paper is found to greatly improve recognition performance under mismatched conditions.
ISBN:9783540308188
3540308180
ISSN:0302-9743
1611-3349
DOI:10.1007/11596448_145