Tobias Herbig, Franz Gerl, Wolfgang Minker - Self-Learning Speaker Identification

《Tobias Herbig, Franz Gerl, Wolfgang Minker - Self-Learning Speaker Identification》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、SignalsandCommunicationTechnologyForfurthervolumes:http://www.springer.com/series/4748TobiasHerbig,FranzGerl,andWolfgangMinkerSelf-LearningSpeakerIdentificationASystemforEnhancedSpeechRecognitionABCAuthorTobiasHerbigWolfgangMinkerInstituteofInformationTechnologyInstituteofInformationTechnologyUnivers

2、ityofUlmUniversityofUlmAlbert-Einstein-Allee43Albert-Einstein-Allee4389081Ulm89081UlmGermanyGermanyE-mail:tobias.herbig@uni-ulm.deE-mail:wolfgang.minker@uni-ulm.deDr.FranzGerlHarmanBeckerAutomotiveSystemsGmbHDept.SpeechRecognitionSöflingerStr.10089077UlmGermanyE-mail:franz.gerl@harman.comISBN978-3-64

3、2-19898-4e-ISBN978-3-642-19899-1DOI10.1007/978-3-642-19899-1LibraryofCongressControlNumber:2011927891c2011Springer-VerlagBerlinHeidelbergThisworkissubjecttocopyright.Allrightsarereserved,whetherthewholeorpartofthemate-rialisconcerned,specificallytherightsoftranslation,reprinting,reuseofillustrations

4、,recitation,broadcasting,reproductiononmicrofilmorinanyotherway,andstorageindatabanks.Dupli-cationofthispublicationorpartsthereofispermittedonlyundertheprovisionsoftheGermanCopyrightLawofSeptember9,1965,initscurrentversion,andpermissionforusemustalwaysbeobtainedfromSpringer.Violationsareliabletoprose

5、cutionundertheGermanCopyrightLaw.Theuseofgeneraldescriptivenames,registerednames,trademarks,etc.inthispublicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.Typeset:ScientificPublishingServicesPvt.Ltd.

6、,Chennai,India.Printedonacid-freepaper987654321springer.comSelf-LearningSpeakerIdentificationforEnhancedSpeechRecognitionSummary.Aself-learningspeechcontrolledsystemhasbeendevelopedforunsupervisedspeakeridentificationandspeechrecognition.Thebenefitsofaspeechcontrolleddevicewhichidentifiesitsmainusersbyt

7、heirvoicechar-acteristicsareobvious:Thehuman-computerinterfacemaybepersonalized.Newwaysforinteractingwithaspeechcontrolledsystemmaybedevelopedtosimplifythehandlingofthedevice.Furthermore,speechrecogni