简介：摘要视频数据分析是监控系统中的重要组成部分，监控系统分析视频的第一步是区分对象，第二步是跟踪对象，最后一步是分析对象行为。本文采用改进的MeanShift算法对行人进行跟踪。

简介：Inthefieldofmedicalimaging,thetraditionallocalbinarypattern(LBP)anditsimprovedalgorithmsareoftensensitivetonoise.TraditionalLBPsaresolelybasedonthesignalinformationfromlocaldifferences,andthebinaryquantizationmethodoversimplifiesthelocaltexturefeatureswhiledisregardingtheimaginginformationfromtheconcaveconvexregionsbetweenthehigh-orderpixelsandtheneighboringsamplingpoints.Therefore,weproposeanimprovedDerivedMeanCompleteLocalBinaryPattern(DM_CLBP)algorithmbasedonhigh-orderderivatives.IntheDM_CLBPmethod,thegreyvalueofasinglepixelisreplacedbythemeangreyvalueoftherectangularareablock,andthedifferencebetweenpixelvaluesintheareaisobtainedusingthesecond-orderdifferentiationmethod.Basedonthecalculationconceptofthecompletelocalbinarypattern(CLBP)algorithm,thecascadesignsandmagnitudesofthetwocomponentsareencodedandrecombinedinDM_CLBPusingauniformpattern.TheresultsfromtheexperimentsshowedthattheproposedDM_CLBPdescriptorsachievedaclassificationaccuracyof94.4%.ComparedwithLBPandotherimprovedalgorithms,theDM_CLBPalgorithmpresentedinthisstudycaneffectivelydifferentiatebetweenlesionareasandnormalareasinthyroidMR,(magneticresonance)imagesandshowstheimprovedaccuracyofareaclassification.

简介：Earthquakeengineershavemadealotofeffortstoderiveacomprehensivesetofclosedformexpressionsforperformanceevaluationofframes,whicharealreadypresentedinguidelinessuchasSAC/FEMA.Theseanalyticalexpressionshavebeendevelopedtoestimatetheannualprobabilityofexceedingalimitstate.Intheprocessofsuchseismicassessments,someessentialassumptionsareadoptedtosimplifytheprocess.Oneofthesefundamentalassumptionsdeclaresthatdriftdemandatanyseismicintensitylevelfollowsalognormaldistributionarounditsmedian.Toinvestigatethevalidityofthisassumption,thispaperdescribesacasestudyofthetypesoferrorsthatcouldbeproducedbyusingthesamplemedianasthecentraltendency.BasedontheMaximumLikelihoodEstimationmethodaswellasotherstatisticalevidence,thispaperproposestheuseofthesamplegeometricmeaninsteadofthesamplemedianforthecentraltendency.Further,theresultsofseismicreliabilityevaluationsof4sampleframesarecomparedbasedonutilizingboththegeometricmeanandthesamplemedian.Inthisprocess,bothfirstandsecondorderpowerlawfitsofthehazardcurveareimplementedtocomparetheeffectsofhazardestimationandtheselectionofthecentraltendencyonthefinalresults.Itisobservedintheapplicationexamplethatthesamplegeometricmeancouldleadtomoreaccurateresults.

简介：Background:Self-efficacyhasbeenidentifiedasanimportantdeterminantofyouth'sbehaviorchangeincludingphysicalactivity(PA)participation.However,thedimensionalitycheckofaPAself-efficacyscalehasrarelybeenconductedinChina.Thecurrentstudyaimstoexamine(1)theunidimensionalityofashortenedChineseversionofPAself-efficacyscale(S-PASESC);(2)themeasurementinvarianceofS-PASESCacrossgenderandlevelsofeducation;(3)thelatentfactormeandifferencebetweengenderandlevelsofeducation;(4)thedirecteffectsofself-efficacyonPAbydifferentgenderandeducationlevels;and(5)thecomparisonsofthedirecteffectsofself-efficacyonPAacrossgenderandeducationlevels.Methods:Theparticipantswere5ththrough11thgradepublicschoolstudentsrecruitedfrom7citieslocatedindifferentgeographicregionsofChina.Thefinaldataincludeatotalof3003participants(49.7%boys)whohavecompletedthescales.Results:Confirmatoryfactoranalysis(CFA)testsupportedtheunidimensionalityofS-PASESC.TheS-PASESCisinvariantacrossgenderand3levelsofeducationatbothconfigural,fullmetric,andfullscalarlevels.FindingsfromlatentmeancomparisonsshowedthatboysreportedhigherPAself-efficacythangirls.Students'perceivedPAself-efficacytendtodecreasefromelementarytohighschool.Finally,self-efficacypositivelyrelatedtoPAbygroupsofdifferentgenderandeducationlevelsandtherelationshipbetweenself-efficacyandPAisstrongeramongmiddleschoolboysthangirls.Conclusion:FindingssuggestS-PASESCisavalidscaleformeasuringChinesestudents'PAself-efficacy.