資源描述：
《改進(jìn)的核函數(shù)算法及其在說(shuō)話人辨認(rèn)中的應(yīng)用研究》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、北京交通大學(xué)碩士學(xué)位論文改進(jìn)的核函數(shù)算法及其在說(shuō)話人辨認(rèn)中的應(yīng)用研究姓名：胡若華申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：通信與信息系統(tǒng)指導(dǎo)教師：張有根20080601ABS瞰CTABSTRACT：Duetoitsspecialmeritsofflexibility,economyandaccuracy,speakerrecognitiontechnologyhasabroadapplicationfutureinbiometricsidentityverificationfield．However,speakerrecognitionhassomelimitsinapplicationbecause
2、thetrainingalgorithmiscomplicated,andtherobustnessisnotideal．SupportVectorMachine(SVM)isanewclassificationmethodology．Ithasbeenprovedtobeapowerfultechniqueinpatternclassificationforitsgoodgeneralizationability．ButSVMhassomedisadvantagesinsomeaspectforit’Sstillinthedevelopingstage．Thethesisfocus
3、onhowtoimprovetherecognitionratioandrobustnessofspeakerrecognitionsystembygeneratingnewkernelsbasedonsupervector．Themaincontributionsofthedissertationareasfollows：(1)Theadvancedfeatureparameterextraction．ThisthesisintroducesGaussianMixtureUniversalBackgroundModel(GMM-UBM)intospeakerrecognitionm
4、odeling．UBMUSeSspeaker-independentdistributionparameterstoapproximateparametersforacousticunitswhichareabsentinspecifiedspeaker’Strainingdata．ThenstackthemeansoftheGMMmodelwhichisadaptedbyMAPalgorithmtoformGMMmeansupervector．(2)Theadoptionofnewkernels，suchasKullbackLeiblerDivergencekernel，rinne
5、rproductkernelandNAPkernel．ThesethreenewkernelsareallbasedonGMMsupervector．TheSVMusingthekernelsbasedonGMMsupervectorCanbeusedtoclassifywhollyonthesequence．Meanwhile，inordertoenhancetherobustnessofthesystem，thethesisadoptkernelbasedonNuisanceAttributeProjection．Thiskindofkernelgetsridofredundan
6、tinformationfromthesubspacewhichhasnorelationshipwiththespeakerfeature．(3)Basedonpeoplevoicedatabase，wetakeemulationexperiments．First,wecomparetheadvancedfeatureextractionmethodwithRBFkernelandpolynomialkernel．Second,weapplythesethreenewkernelstospeakeridentification．Fromtheresult，wecanseethese
7、threenewkernelsimprovedrecognitionratioatleastby12％，andNAPkernelimprovedtherobustnessofthesystemalot．KEYWORDS：Speakeridentification；SupportVectorMachine；GMMSupervectorKernel；PrincipalComponentAnalysis(PCA)；NuisanceAttributeProject
