表情分析技术在可用性评估中的应用研究——以facereader软件为例

《表情分析技术在可用性评估中的应用研究——以facereader软件为例》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、万方数据浙江理工大学硕士学位论文验结果，相对于改进后可用性水平较高的产品，可用性水平较差的改进前产品不但操作时间更长，正确率更低，负性表情成分更多，SUS评分更低，且在表情强度上，SUS评分与“悲伤”、“生气”、“惊奇”等负性情绪成分问存在显著负相关。而正性表情成分指标相对来说没有显著差异。关键词：表情识别；FaceReader；有效性；可用性万方数据浙江理工大学硕士学位论文AbstractFacialexpressionanalysismethodinusabilityevaluationreferstodirectmeasureandana

2、lysisoftheexpressionsgeneratedintheprocessofusingtheproduct．Withthedirectandreal-timeadvantage，facialexpressionanalysismakesupforthetraditionalusabilitymethods’indirect,hystereticdisadvantage．Hence，moreandmoreresearchersbegantOpayattentiontothevalueoftheexpressionanalysis．Th

3、isstudytookFaceReadersoftwareasanexample，firstly,thefacialexpressionrecognitioneffectivenessinChinesewasexaminedbymeansofstandardexpressionsdatabase，andthentheapplicationofexpressionindexinusabilityevaluationWasexploredusingFaceReaderasat001．Thisresearchprovidedascientificba

4、sisforusingexpressionanalysismethodinproductusabilityevaluation．ThispaperincludedtWOstudies．Thedetailswereasfollows：Studyl：InvestigateeffectivenessofFaceReaderexpressionanalysissoftwareonthefacialexpressionrecognitioninChina．ThispartofthestudyusedthreekindsofimagefromUSTC-NV

5、IEstandardexpressiondatabase：artificialfacialemotion，spontaneousfacialexpression，dynamicexpressiontriggeredbystandardemotionstimulation．BycomparingrecognitionrateandStrengthconsistencybetWeenFaceReaderexpressionanalysissoftwareandhumanevaluatorsofthreekindsofexpressionstimul

6、ation，thestudyinvestigatedeffectivenessofFaceReaderexpressionanalysissoftwareonthefacialexpressionrecognitioninChina．Firstly,forartificialfacialexpressionimagesandspontaneousfacialexpressionpictures，standardizedselectwasfirstlyconducted．Thentherecognitionrateandstrengthconsi

7、stencyfromFaceReaderandevaluatorswascomparedindifferentexpressiontypes．Secondly,forfacialexpressionvideos，standardizedemotionelicitingstimulusWasusedtocollectexpressionvideomaterial，thenrecognitionrateandintensityconsistencyWascomparedfromFaceReaderandsubjects’self-report,Fa

8、ceReaderandevaluatorsindifferentexpressiontypes．万方数据浙江理工大学硕士学位论文Study2：Appl