aaas00-dar-pres

aaas00-dar-pres

ID:33870334

大小:3.57 MB

页数:36页

时间:2019-02-28

aaas00-dar-pres_第1页
aaas00-dar-pres_第2页
aaas00-dar-pres_第3页
aaas00-dar-pres_第4页
aaas00-dar-pres_第5页
资源描述:

《aaas00-dar-pres》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库

1、AutomaticSpeakerRecognitionRecentProgress,CurrentApplications,andFutureTrendsDouglasA.Reynolds,PhDLarryP.Heck,PhDSeniorMemberofTechnicalStaffManager,SpeakerVerificationR&DM.I.T.LincolnLaboratoryNuanceCommunicationsPresentedattheAAAS2000MeetingHumans,

2、ComputersandSpeechSymposium19February2000ThisworkwassponsoredbytheDepartmentofDefenseunderAirForcecontractF19628-95-C-0002.Opinions,interpretations,conclusions,andrecommendationsarethoseoftheauthorsandarenotnecessarilyendorsedbytheUnitedStatesAirForc

3、e.NuanceCommunicationsMITLincolnLaboratoryOutline•Introduction(Reynolds)•Generaltheory(Reynolds)•Performance(Heck)•Applications(Heck)•Conclusionsandfuturedirections(Heck)NuanceCommunicationsMITLincolnLaboratoryExtractingInformationfromSpeechGoal:Auto

4、maticallyextractinformationtransmittedinspeechsignalSpeechWordsRecognition“Howareyou?”LanguageLanguageNameRecognitionSpeechSignalEnglishSpeakerSpeakerNameRecognitionJamesWilsonNuanceCommunicationsMITLincolnLaboratoryIntroductionIdentification•Determi

5、neswhoistalkingfromsetofknownvoices•Noidentityclaimfromuser(manytoonemapping)•Oftenassumedthatunknownvoicemustcomefromsetofknownspeakers-referredtoasclosed-setidentification?Whosevoiceisthis????NuanceCommunicationsMITLincolnLaboratoryIntroductionVeri

6、fication/Authentication/Detection•Determinewhetherpersoniswhohe/sheclaimstobe•Usermakesidentityclaim:onetoonemapping•Unknownvoicecouldcomefromlargesetofunknownspeakers-referredtoasopen-setverification•Adding“none-of-the-above”optiontoclosed-setidenti

7、ficationgivesopen-setidentificationIsthisBob’svoice??NuanceCommunicationsMITLincolnLaboratoryIntroductionSpeechModalitiesApplicationdictatesdifferentspeechmodalities:•Text-dependentrecognition–Recognitionsystemknowstextspokenbyperson–Examples:fixedph

8、rase,promptedphrase–Usedforapplicationswithstrongcontroloveruserinput–Knowledgeofspokentextcanimprovesystemperformance•Text-independentrecognition–Recognitionsystemdoesnotknowtextspokenbyperson–Examples:Userselectedphrase,conversationalspeech–Usedfor

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。
相关文章
更多
相关标签