A Study of the Automatic Speech Recognition Process and Speaker Adaptation

Simple item page
dc.contributor.authorStokes-Rees, Ian Jamesen
dc.date.accessioned2006-08-22T13:57:36Z
dc.date.available2006-08-22T13:57:36Z
dc.date.issued2000en
dc.date.submitted2000en
dc.description.abstractThis thesis considers the entire automated speech recognition process and presents a standardised approach to LVCSR experimentation with HMMs. It also discusses various approaches to speaker adaptation such as MLLR and multiscale, and presents experimental results for cross­-task speaker adaptation. An analysis of training parameters and data sufficiency for reasonable system performance estimates are also included. It is found that Maximum Likelihood Linear Regression (MLLR) supervised adaptation can result in 6% reduction (absolute) in word error rate given only one minute of adaptation data, as compared with an unadapted model set trained on a different task. The unadapted system performed at 24% WER and the adapted system at 18% WER. This is achieved with only 4 to 7 adaptation classes per speaker, as generated from a regression tree.en
dc.formatapplication/pdfen
dc.format.extent512540 bytes
dc.format.mimetypeapplication/pdf
dc.identifier.urihttp://hdl.handle.net/10012/840
dc.language.isoenen
dc.pendingfalseen
dc.publisherUniversity of Waterlooen
dc.rightsCopyright: 2000, Stokes-Rees, Ian James. All rights reserved.en
dc.subjectElectrical & Computer Engineeringen
dc.subjectautomatic speech recognitionen
dc.subjectspeaker adaptationen
dc.subjectHTKen
dc.subjectHMMen
dc.subjectMLLRen
dc.subjectLVCSRen
dc.titleA Study of the Automatic Speech Recognition Process and Speaker Adaptationen
dc.typeMaster Thesisen
uws-etd.degreeMaster of Applied Scienceen
uws-etd.degree.departmentElectrical and Computer Engineeringen
uws.peerReviewStatusUnrevieweden
uws.scholarLevelGraduateen
uws.typeOfResourceTexten

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
ijstokes2000.pdf
Size:
500.53 KB
Format:
Adobe Portable Document Format
Download

Collections

Theses
Electrical and Computer Engineering