欢迎来到天天文库
浏览记录
ID:33991128
大小:7.74 MB
页数:55页
时间:2019-03-02
《基于脑电信号的喜好度分析》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、上海交通大学硕士学位论文基于脑电信号的喜好度分析姓名:傅佳伟申请学位级别:硕士专业:计算机软件与理论指导教师:吕宝粮20090101ÄuM>&ÒUÐÝ©ÛÁX´«aú!gÚ1'«°Ú%nþG"§
2、XÌ*Naɧ33、ÄuÀúpu@¢§4、4^r*wk¡L&E<ò졧ÓP¹e^rM>&Ò§¦^ÅìÆS{5©aù¹XA&EM>&Ò§5±OÑ^rUÐÝG8"3êâ?nL§¥§·¦^~^uM>êâ?nmª{5JA&Ò§¿¦^5、±þÅ{éAêâ?1©a§¿ÄudJÑ«dµeeVÇmª{" ©a(JL²úX±ÏLM>&Ò«©m§¿Ñ)nÆ)º"© zXeµÄk§·O6、ÄuÀúpu@¢§æ8½êþ¹X&EM>ê⧿éêâ?1'ÜnI5§¦^ÅìÆS{ÏLM>&Ò57、O^rUÐݶÙg§·¦^mª{§éM>&Ò?1?n§T{±k/Ø©a8Ã'µ&Ò"1n§·JÑ«dµeeVÇmª{§é¢Â8êâ?1?n§©a°Ý½§UÐ/;N´L[ܹ"'cµM>&Ò§UÐݧmª§VÇmª§¯'>IValenceAnalysisBasedonEEGSignalsABSTRACTEmotionisamentalandpsychologicalstaterelatedtovariousfeelings,thoughtsandbehavio8、rs.Itdominatessubject'sexperienceandplaysanimportantroleinhuman'sdailylife.Itisavitalbasisforthedevelopmentofintellectualsystemtoenablethemachinetofeeltheuser'semotion.Thiscanmakeanintellectualmachinemorehumanistic.Anditcanimprovetheuser'sexperience,maket9、heusercommunicatemorenaturallywiththesystem.Valenceisameasurementwhichisoftenusedtoevaluateemotioninpsychology.Anditcanbeusedtolabeldifferentemotionstate,e.g.joyandsadnessaretotallydifferentinvalence.Wetrytoestimatetheuser'svalencebycarryingoutavisual-evo10、kedrecognitionexperiment.Westimulatedthesubjectbyfacialpictureswithjoyfulandsadexpression,andrecordedtheEEGmeanwhile.Byapplyingsomemethodsofmachinelearning,wetrytodiffertheseEEGsignalsintojoyandsadnesstoestimatetheuser'svalence.Whenprocessingthedata,weapp11、lythetraditionalcommonspatialpatterntoextracttheusefulfeatures.Andsupportvectormachineisusedasclassifier.AnovelalgorithmbasedonBayesianisproposedhere,whichiscalledprobabilitycommonspatialpattern.TheresultshowsthatjoyandsadnesscanbedistinguishedbyusingEEG.A12、ndwegavesomepsychologicalexplanationatlast.Thecontributionsofthispaperareasfollows:Firstly,wedesignedavisual-evokedrecognitionexperiment.SomeEEGdatawithemotioninformationwererecorded.WelabeledthesedataaccordingtoSAM
3、ÄuÀúpu@¢§
4、4^r*wk¡L&E<ò졧ÓP¹e^rM>&Ò§¦^ÅìÆS{5©aù¹XA&EM>&Ò§5±OÑ^rUÐÝG8"3êâ?nL§¥§·¦^~^uM>êâ?nmª{5JA&Ò§¿¦^
5、±þÅ{éAêâ?1©a§¿ÄudJÑ«dµeeVÇmª{" ©a(JL²úX±ÏLM>&Ò«©m§¿Ñ)nÆ)º"© zXeµÄk§·O
6、ÄuÀúpu@¢§æ8½êþ¹X&EM>ê⧿éêâ?1'ÜnI5§¦^ÅìÆS{ÏLM>&Ò5
7、O^rUÐݶÙg§·¦^mª{§éM>&Ò?1?n§T{±k/Ø©a8Ã'µ&Ò"1n§·JÑ«dµeeVÇmª{§é¢Â8êâ?1?n§©a°Ý½§UÐ/;N´L[ܹ"'cµM>&Ò§UÐݧmª§VÇmª§¯'>IValenceAnalysisBasedonEEGSignalsABSTRACTEmotionisamentalandpsychologicalstaterelatedtovariousfeelings,thoughtsandbehavio
8、rs.Itdominatessubject'sexperienceandplaysanimportantroleinhuman'sdailylife.Itisavitalbasisforthedevelopmentofintellectualsystemtoenablethemachinetofeeltheuser'semotion.Thiscanmakeanintellectualmachinemorehumanistic.Anditcanimprovetheuser'sexperience,maket
9、heusercommunicatemorenaturallywiththesystem.Valenceisameasurementwhichisoftenusedtoevaluateemotioninpsychology.Anditcanbeusedtolabeldifferentemotionstate,e.g.joyandsadnessaretotallydifferentinvalence.Wetrytoestimatetheuser'svalencebycarryingoutavisual-evo
10、kedrecognitionexperiment.Westimulatedthesubjectbyfacialpictureswithjoyfulandsadexpression,andrecordedtheEEGmeanwhile.Byapplyingsomemethodsofmachinelearning,wetrytodiffertheseEEGsignalsintojoyandsadnesstoestimatetheuser'svalence.Whenprocessingthedata,weapp
11、lythetraditionalcommonspatialpatterntoextracttheusefulfeatures.Andsupportvectormachineisusedasclassifier.AnovelalgorithmbasedonBayesianisproposedhere,whichiscalledprobabilitycommonspatialpattern.TheresultshowsthatjoyandsadnesscanbedistinguishedbyusingEEG.A
12、ndwegavesomepsychologicalexplanationatlast.Thecontributionsofthispaperareasfollows:Firstly,wedesignedavisual-evokedrecognitionexperiment.SomeEEGdatawithemotioninformationwererecorded.WelabeledthesedataaccordingtoSAM
此文档下载收益归作者所有