Uncertainty Decoding for Reverberation-Robust Automatic Speech Recognition

The major problem in distant-talking speech recognition is the corruption of speech signals by both interfering sounds and reverberation. While a range of successful techniques has been developed since the beginnings of speech recognition research to combat additive and short convolutive noise, comp...

Olles dieđut

Furkejuvvon:
Bibliográfalaš dieđut
Váldodahkki: Maas, Roland
Materiálatiipa: Online
Giella:eaŋgalasgiella
Almmustuhtton: FAU University Press 2025
Fáttát:
Liŋkkat:ONIX_20250828T094736_9783944057620_38
Fáddágilkorat: Lasit fáddágilkoriid
Eai fáddágilkorat, Lasit vuosttaš fáddágilkora!
_version_ 1869517169072537600
author Maas, Roland
author_browse Maas, Roland
author_facet Maas, Roland
author_sort Maas, Roland
collection Directory of Open Access Books
description The major problem in distant-talking speech recognition is the corruption of speech signals by both interfering sounds and reverberation. While a range of successful techniques has been developed since the beginnings of speech recognition research to combat additive and short convolutive noise, compensating for long-term distortion caused by reverberation has not gained wide attention until recently. This thesis further develops an uncertainty decoding approach, named REverberation MOdeling for Speech recognition (REMOS), to adapt the acoustic model of a conventional Hidden Markov Model-based recognizer to reverberant environments. By incorporating a convolutive observation model, the Viterbi decoder is extended in order to implicitly provide a state-wise late reverberation estimate leading to a relaxation of the hidden Markov models' conditional independence assumption. The experimental evaluation confirms that REMOS yields strong speech recognition performance under noisy and reverberant conditions and furthermore allows for a rapid adaptation to changing acoustic conditions.
format Online
id doab-20.500.12854ir-166270
institution Directory of Open Access Books
language eng
publishDate 2025
publishDateRange 2025
publishDateSort 2025
publisher FAU University Press
publisherStr FAU University Press
record_format ojs
spelling doab-20.500.12854ir-1662702025-10-16T13:06:45Z Uncertainty Decoding for Reverberation-Robust Automatic Speech Recognition Maas, Roland Automatische Spracherkennung Hidden-Markov-Modell Nachhall thema EDItEUR::T Technology, Engineering, Agriculture, Industrial processes The major problem in distant-talking speech recognition is the corruption of speech signals by both interfering sounds and reverberation. While a range of successful techniques has been developed since the beginnings of speech recognition research to combat additive and short convolutive noise, compensating for long-term distortion caused by reverberation has not gained wide attention until recently. This thesis further develops an uncertainty decoding approach, named REverberation MOdeling for Speech recognition (REMOS), to adapt the acoustic model of a conventional Hidden Markov Model-based recognizer to reverberant environments. By incorporating a convolutive observation model, the Viterbi decoder is extended in order to implicitly provide a state-wise late reverberation estimate leading to a relaxation of the hidden Markov models' conditional independence assumption. The experimental evaluation confirms that REMOS yields strong speech recognition performance under noisy and reverberant conditions and furthermore allows for a rapid adaptation to changing acoustic conditions. 2025-08-29T05:08:23Z 2025-08-29T05:08:23Z 2025-08-28T08:00:05Z 2016 book ONIX_20250828T094736_9783944057620_38 https://library.oapen.org/handle/20.500.12657/105794 9783944057620 9783944057613 https://directory.doabooks.org/handle/20.500.12854/166270 eng FAU Forschungen : Reihe B open access image/jpeg image/jpeg n/a n/a https://library.oapen.org/bitstream/20.500.12657/105794/1/9783944057620.pdf https://library.oapen.org/bitstream/20.500.12657/105794/1/9783944057620.pdf FAU University Press 2c600dea-eece-4066-87be-da335e323fdb 9783944057620 9783944057613 AG Universitätsverlage 191 Erlangen open access
spellingShingle Automatische Spracherkennung
Hidden-Markov-Modell
Nachhall
thema EDItEUR::T Technology, Engineering, Agriculture, Industrial processes
Maas, Roland
Uncertainty Decoding for Reverberation-Robust Automatic Speech Recognition
title Uncertainty Decoding for Reverberation-Robust Automatic Speech Recognition
title_full Uncertainty Decoding for Reverberation-Robust Automatic Speech Recognition
title_fullStr Uncertainty Decoding for Reverberation-Robust Automatic Speech Recognition
title_full_unstemmed Uncertainty Decoding for Reverberation-Robust Automatic Speech Recognition
title_short Uncertainty Decoding for Reverberation-Robust Automatic Speech Recognition
title_sort uncertainty decoding for reverberation robust automatic speech recognition
topic Automatische Spracherkennung
Hidden-Markov-Modell
Nachhall
thema EDItEUR::T Technology, Engineering, Agriculture, Industrial processes
topic_facet Automatische Spracherkennung
Hidden-Markov-Modell
Nachhall
thema EDItEUR::T Technology, Engineering, Agriculture, Industrial processes
url ONIX_20250828T094736_9783944057620_38
work_keys_str_mv AT maasroland uncertaintydecodingforreverberationrobustautomaticspeechrecognition