schliessen

Filtern

 

Bibliotheken

Adaptive bands filter bank optimized by genetic algorithm for robust speech recognition system

Perceptual auditory filter banks such as Bark-scale filter bank are widely used as front-end processing in speech recognition systems. However, the problem of the design of optimized filter banks that provide higher accuracy in recognition tasks is still open. Owing to spectral analysis in feature e... Full description

Journal Title: Journal of Central South University of Technology 2011, Vol.18(5), pp.1595-1601
Main Author: Huang, Li-xia
Other Authors: Evangelista, G. , Zhang, Xue-ying
Format: Electronic Article Electronic Article
Language: English
Subjects:
ID: ISSN: 1005-9784 ; E-ISSN: 1993-0666 ; DOI: 10.1007/s11771-011-0877-1
Link: http://dx.doi.org/10.1007/s11771-011-0877-1
Zum Text:
SendSend as email Add to Book BagAdd to Book Bag
Staff View
recordid: springer_jour10.1007/s11771-011-0877-1
title: Adaptive bands filter bank optimized by genetic algorithm for robust speech recognition system
format: Article
creator:
  • Huang, Li-xia
  • Evangelista, G.
  • Zhang, Xue-ying
subjects:
  • perceptual filter banks
  • bark scale
  • speaker independent speech recognition systems
  • zero-crossing peak amplitude
  • genetic algorithm
ispartof: Journal of Central South University of Technology, 2011, Vol.18(5), pp.1595-1601
description: Perceptual auditory filter banks such as Bark-scale filter bank are widely used as front-end processing in speech recognition systems. However, the problem of the design of optimized filter banks that provide higher accuracy in recognition tasks is still open. Owing to spectral analysis in feature extraction, an adaptive bands filter bank (ABFB) is presented. The design adopts flexible bandwidths and center frequencies for the frequency responses of the filters and utilizes genetic algorithm (GA) to optimize the design parameters. The optimization process is realized by combining the front-end filter bank with the back-end recognition network in the performance evaluation loop. The deployment of ABFB together with zero-crossing peak amplitude (ZCPA) feature as a front process for radial basis function (RBF) system shows significant improvement in robustness compared with the Bark-scale filter bank. In ABFB, several sub-bands are still more concentrated toward lower frequency but their exact locations are determined by the performance rather than the perceptual criteria. For the ease of optimization, only symmetrical bands are considered here, which still provide satisfactory results.
language: eng
source:
identifier: ISSN: 1005-9784 ; E-ISSN: 1993-0666 ; DOI: 10.1007/s11771-011-0877-1
fulltext: fulltext
issn:
  • 1993-0666
  • 19930666
  • 1005-9784
  • 10059784
url: Link


@attributes
ID170915056
RANK0.07
NO1
SEARCH_ENGINEprimo_central_multiple_fe
SEARCH_ENGINE_TYPEPrimo Central Search Engine
LOCALfalse
PrimoNMBib
record
control
sourcerecordid10.1007/s11771-011-0877-1
sourceidspringer_jour
recordidTN_springer_jour10.1007/s11771-011-0877-1
sourcesystemPC
galeid268792764
display
typearticle
titleAdaptive bands filter bank optimized by genetic algorithm for robust speech recognition system
creatorHuang, Li-xia ; Evangelista, G. ; Zhang, Xue-ying
ispartofJournal of Central South University of Technology, 2011, Vol.18(5), pp.1595-1601
identifier
subjectperceptual filter banks ; bark scale ; speaker independent speech recognition systems ; zero-crossing peak amplitude ; genetic algorithm
descriptionPerceptual auditory filter banks such as Bark-scale filter bank are widely used as front-end processing in speech recognition systems. However, the problem of the design of optimized filter banks that provide higher accuracy in recognition tasks is still open. Owing to spectral analysis in feature extraction, an adaptive bands filter bank (ABFB) is presented. The design adopts flexible bandwidths and center frequencies for the frequency responses of the filters and utilizes genetic algorithm (GA) to optimize the design parameters. The optimization process is realized by combining the front-end filter bank with the back-end recognition network in the performance evaluation loop. The deployment of ABFB together with zero-crossing peak amplitude (ZCPA) feature as a front process for radial basis function (RBF) system shows significant improvement in robustness compared with the Bark-scale filter bank. In ABFB, several sub-bands are still more concentrated toward lower frequency but their exact locations are determined by the performance rather than the perceptual criteria. For the ease of optimization, only symmetrical bands are considered here, which still provide satisfactory results.
languageeng
source
version6
lds50peer_reviewed
links
openurl$$Topenurl_article
openurlfulltext$$Topenurlfull_article
backlink$$Uhttp://dx.doi.org/10.1007/s11771-011-0877-1$$EView_full_text_in_Springer_(Subscribers_only)
search
creatorcontrib
0Huang, Li-xia
1Evangelista, G.
2Zhang, Xue-ying
titleAdaptive bands filter bank optimized by genetic algorithm for robust speech recognition system
descriptionPerceptual auditory filter banks such as Bark-scale filter bank are widely used as front-end processing in speech recognition systems. However, the problem of the design of optimized filter banks that provide higher accuracy in recognition tasks is still open. Owing to spectral analysis in feature extraction, an adaptive bands filter bank (ABFB) is presented. The design adopts flexible bandwidths and center frequencies for the frequency responses of the filters and utilizes genetic algorithm (GA) to optimize the design parameters. The optimization process is realized by combining the front-end filter bank with the back-end recognition network in the performance evaluation loop. The deployment of ABFB together with zero-crossing peak amplitude (ZCPA) feature as a front process for radial basis function (RBF) system shows significant improvement in robustness compared with the Bark-scale filter bank. In ABFB, several sub-bands are still more concentrated toward lower frequency but their exact locations are determined by the performance rather than the perceptual criteria. For the ease of optimization, only symmetrical bands are considered here, which still provide satisfactory results.
subject
0perceptual filter banks
1bark scale
2speaker independent speech recognition systems
3zero-crossing peak amplitude
4genetic algorithm
general
010.1007/s11771-011-0877-1
1English
2Springer Science & Business Media B.V.
3SpringerLink
sourceidspringer_jour
recordidspringer_jour10.1007/s11771-011-0877-1
issn
01993-0666
119930666
21005-9784
310059784
rsrctypearticle
creationdate2011
addtitle
0Journal of Central South University of Technology
1Science & Technology of Mining and Metallurgy
2J. Cent. South Univ. Technol.
searchscopespringer_journals_complete
scopespringer_journals_complete
lsr30VSR-Enriched:[pages, eissn, galeid]
sort
titleAdaptive bands filter bank optimized by genetic algorithm for robust speech recognition system
authorHuang, Li-xia ; Evangelista, G. ; Zhang, Xue-ying
creationdate20111000
facets
frbrgroupid1763177471988014626
frbrtype5
languageeng
creationdate2011
topic
0Perceptual Filter Banks
1Bark Scale
2Speaker Independent Speech Recognition Systems
3Zero-Crossing Peak Amplitude
4Genetic Algorithm
collectionSpringerLink
prefilterarticles
rsrctypearticles
creatorcontrib
0Huang, Li-xia
1Evangelista, G.
2Zhang, Xue-ying
jtitleJournal Of Central South University Of Technology
toplevelpeer_reviewed
delivery
delcategoryRemote Search Resource
fulltextfulltext
addata
aulast
0Huang
1Evangelista
2Zhang
aufirst
0Li-xia
1G.
2Xue-ying
au
0Huang, Li-xia
1Evangelista, G.
2Zhang, Xue-ying
atitleAdaptive bands filter bank optimized by genetic algorithm for robust speech recognition system
jtitleJournal of Central South University of Technology
stitleJ. Cent. South Univ. Technol.
addtitleScience & Technology of Mining and Metallurgy
risdate201110
volume18
issue5
spage1595
epage1601
issn1005-9784
eissn
01993-0666
122275223
genrearticle
ristypeJOUR
abstractPerceptual auditory filter banks such as Bark-scale filter bank are widely used as front-end processing in speech recognition systems. However, the problem of the design of optimized filter banks that provide higher accuracy in recognition tasks is still open. Owing to spectral analysis in feature extraction, an adaptive bands filter bank (ABFB) is presented. The design adopts flexible bandwidths and center frequencies for the frequency responses of the filters and utilizes genetic algorithm (GA) to optimize the design parameters. The optimization process is realized by combining the front-end filter bank with the back-end recognition network in the performance evaluation loop. The deployment of ABFB together with zero-crossing peak amplitude (ZCPA) feature as a front process for radial basis function (RBF) system shows significant improvement in robustness compared with the Bark-scale filter bank. In ABFB, several sub-bands are still more concentrated toward lower frequency but their exact locations are determined by the performance rather than the perceptual criteria. For the ease of optimization, only symmetrical bands are considered here, which still provide satisfactory results.
copHeidelberg
pubCentral South University
doi10.1007/s11771-011-0877-1
pages1595-1601
date2011-10