A New Evidence Model for Missing Data Speech Recognition With Applications in Reverberant Multi-Source Environments

Conventional hidden Markov model (HMM) decoders often experience severe performance degradations in practice due to their inability to cope with uncertain data in time-varying environments. In order to address this issue, we propose the bounded-Gauss-Uniform mixture probablity density function (pdf)...

Full description

Bibliographic Details
Main Authors: Kuhne, M., Togneri, R., Nordholm, Sven
Format: Journal Article
Published: IEEE Signal Processing Society 2011
Online Access:http://hdl.handle.net/20.500.11937/36504
_version_ 1848754788851777536
author Kuhne, M.
Togneri, R.
Nordholm, Sven
author_facet Kuhne, M.
Togneri, R.
Nordholm, Sven
author_sort Kuhne, M.
building Curtin Institutional Repository
collection Online Access
description Conventional hidden Markov model (HMM) decoders often experience severe performance degradations in practice due to their inability to cope with uncertain data in time-varying environments. In order to address this issue, we propose the bounded-Gauss-Uniform mixture probablity density function (pdf) as a new class of evidence model for missing data speech recognition. Exemplary for a hands-free speech recognition scenario, we illustrate how the parameters of the new mixture pdf can be estimated with the help of a multi-channel source separation front-ed. In comparison with other models the new evidence pdf retains a fuller description of the available data and provides a more effective link between source separation and recognition. The superiority of the bounded-Gauss-Uniform mixture pdf over conventional approaches is demonstrated for a connected digits recognition task under varying test conditions.
first_indexed 2025-11-14T08:45:59Z
format Journal Article
id curtin-20.500.11937-36504
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T08:45:59Z
publishDate 2011
publisher IEEE Signal Processing Society
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-365042017-09-13T16:08:47Z A New Evidence Model for Missing Data Speech Recognition With Applications in Reverberant Multi-Source Environments Kuhne, M. Togneri, R. Nordholm, Sven Conventional hidden Markov model (HMM) decoders often experience severe performance degradations in practice due to their inability to cope with uncertain data in time-varying environments. In order to address this issue, we propose the bounded-Gauss-Uniform mixture probablity density function (pdf) as a new class of evidence model for missing data speech recognition. Exemplary for a hands-free speech recognition scenario, we illustrate how the parameters of the new mixture pdf can be estimated with the help of a multi-channel source separation front-ed. In comparison with other models the new evidence pdf retains a fuller description of the available data and provides a more effective link between source separation and recognition. The superiority of the bounded-Gauss-Uniform mixture pdf over conventional approaches is demonstrated for a connected digits recognition task under varying test conditions. 2011 Journal Article http://hdl.handle.net/20.500.11937/36504 10.1109/TASL.2010.2048604 IEEE Signal Processing Society restricted
spellingShingle Kuhne, M.
Togneri, R.
Nordholm, Sven
A New Evidence Model for Missing Data Speech Recognition With Applications in Reverberant Multi-Source Environments
title A New Evidence Model for Missing Data Speech Recognition With Applications in Reverberant Multi-Source Environments
title_full A New Evidence Model for Missing Data Speech Recognition With Applications in Reverberant Multi-Source Environments
title_fullStr A New Evidence Model for Missing Data Speech Recognition With Applications in Reverberant Multi-Source Environments
title_full_unstemmed A New Evidence Model for Missing Data Speech Recognition With Applications in Reverberant Multi-Source Environments
title_short A New Evidence Model for Missing Data Speech Recognition With Applications in Reverberant Multi-Source Environments
title_sort new evidence model for missing data speech recognition with applications in reverberant multi-source environments
url http://hdl.handle.net/20.500.11937/36504