基于雙層分類模型的實時語音情感與音素識別.pdf

1、浙江大學(xué)計算機(jī)科學(xué)與技術(shù)學(xué)院碩士學(xué)位論文基于雙層分類模型的實時語音情感與音素識別姓名：王博申請學(xué)位級別：碩士專業(yè)：計算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：卜佳俊20080501浙江大學(xué)碩士學(xué)位論文AbslractAbstractInteractionbyspeechisthemostnaturalwaytocommunicate謝mcomputer，asthebasisforit，speechrecognitionhasalwaysbeenahotre

2、searchareaspeechrecognitionincludesspeechsignalpre—processing，acousticfeaturesextraction，dimensionalityreductionandstatisticalmodelb藍(lán)edspeechrecognitionThethesisfocusontwomajorfacetsofspeechrecognitionemotionrecognitiona

3、ndphonemerecognitionThroughtheimprovementofexistingstatisticalmodels，wefmaUyachieveottrgoal：realtimerecognitionofemotionandphonemeFirstofalt，thebackgroundandcontentalegivenThentheimportanceofemotionandphonemerecognition，

4、acousticfeatures，statisticalmoddsandalgorithmsofdimensionalityreductionaregivenAdaBoostC45，whichiswidelyusedintheareasoftextclassificationimagerecognitionandSOonisintroducedintospeechareaandimprovedtooutputtwolabelsUs噸th

5、isnewmodel，thegoatofreal—timeemotionandphonemerecognitionis丘I】_aayachievedOnthefirstlevelofemotionrecognition，inordertohighlightthecontinuityofemotionthedifferenceofacousticfeaturesbetweentwoadjacentframesisintroduced，in

6、ordertoavoidthejitterphenomenonofemotion，aweightedvotinggrategyisused；Onthesecondlevelofphonemerecognition，inordertoavoidcoarticulationaformerflameasSOciatedspeechmodelisdevelopedandtheIsomapalgorithmisalsousedtosavetime

7、Themodeldescribedinthisthesis，canbeexploitedinmanyapplicationssuchasvideoconferencing，games，virtual“showmala“onlinechattingandvirtualreality，etcEmotionRecognition，PhonemeRecognition，C4，5，AdaBoostDifferenceofAcousticFeatu