Dear Spiral user, We’re pleased to introduce Helix, https://helix.imperial.ac.uk/ Imperial’s new repository for research datasets. Please use Helix for all future dataset deposits. There is no change to Spiral - continue depositing as usual. For queries, contact the Research Data Management team at rdm-enquiries@imperial.ac.uk. Kind regards, the Spiral Team. The Spiral team
Reverberant speech recognition exploiting clarity index estimation
File(s)
Author(s)
Parada, PP
Sharma, D
Naylor, PA
van Waterschoot, T
Type
Journal Article
Abstract
We present single-channel approaches to robust automatic speech recognition (ASR) in reverberant environments based on non-intrusive estimation of the clarity index (C 50). Our best performing method includes the estimated value of C 50 in the ASR feature vector and also uses C 50 to select the most suitable ASR acoustic model according to the reverberation level. We evaluate our method on the REVERB Challenge database employing two different C 50 estimators and show that our method outperforms the best baseline of the challenge achieved without unsupervised acoustic model adaptation, i.e. using multi-condition hidden Markov models (HMMs). Our approach achieves a 22.4 % relative word error rate reduction in comparison to the best baseline of the challenge.
Date Issued
2015-07-01
Date Acceptance
2015-06-05
Citation
Eurasip Journal on Advances in Signal Processing, 2015, 2016, pp.1-12
URI
DOI
ISSN
1687-6180
Publisher
Hindawi Publishing Corporation
Start Page
1
End Page
12
Journal / Book Title
Eurasip Journal on Advances in Signal Processing
Volume
2016
Copyright Statement
© Parada et al. 2015. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://​creativecommons.​org/​licenses/​by/​4.​0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.
Sponsor
Commission of the European Communities
Identifier
http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcApp=PARTNER_APP&SrcAuth=LinksAMR&KeyUT=WOS:000358321400001&DestLinkType=FullRecord&DestApp=ALL_WOS&UsrCustomerID=1ba7043ffcc86c417c072aa74d649202
Grant Number
PITN-GA-2012-316969
Subjects
Science & Technology
Technology
Engineering, Electrical & Electronic
Engineering
Reverberant speech recognition
C-50
Acoustic model selection
Networking & Telecommunications
Electrical And Electronic Engineering
Artificial Intelligence And Image Processing
Publication Status
Published
Article Number
ARTN 54