简介:Analgorithmforfacedescriptionandrecognitionbasedonmulti-resolutionwithmulti-scalelocalbinarypattern(multi-LBP)featuresisproposed.Thefacialimagepyramidisconstructedandeachfacialimageisdividedintovariousregionsfromwhichpartialandholisticlocalbinarypatter(LBP)histogramsareextracted.AllLBPfeaturesofeachimageareconcatenatedtoasingleLBPeigenvectorwithdifferentresolutions.ThedimensionalityofLBPfeaturesisthenreducedbyalocalmarginalignment(LMA)algorithmbasedonmanifold,whichcanpreservethebetween-classvariance.Supportvectormachine(SVM)isappliedtoclassifyfacialimages.ExtensiveexperimentsonORLandCMUfacedatabasesclearlyshowthesuperiorityoftheproposedschemeoversomeexistedalgorithms,especiallyontherobustnessofthemethodagainstdifferentfacialexpressionsandposturesofthesubjects.
简介:Inordertosolvetheproblemoflowrecognitionrateoftraditionalfeatureextractionoperatorsunderlow-resolutionimages,anovelalgorithmofexpressionrecognitionisproposed,namedcentralobliqueaveragecenter-symmetriclocalbinarypattern(CS-LBP)withadaptivethreshold(ATCS-LBP).Firstly,thefeaturesoffaceimagescanbeextractedbytheproposedoperatorafterpretreatment.Secondly,theobtainedfeatureimageisdividedintoblocks.Thirdly,thehistogramofeachblockiscomputedindependentlyandallhistogramscanbeconnectedseriallytocreateafinalfeaturevector.Finally,expressionclassificationisachievedbyusingsupportvectormachine(SVM)classifier.ExperimentalresultsonJapanesefemalefacialexpression(JAFFE)databaseshowthattheproposedalgorithmcanachievearecognitionrateof81.9%whentheresolutionisaslowas16×16,whichismuchbetterthanthatofthetraditionalfeatureextractionoperators.
简介:Inthispaper,wedemonstratetheacetylehe(C2H2)sensorwithhighsensitivityusingahollow-corephotonicbandgapfiber(HC-PCF).ExperimentsformeasuringC2H2concentrationsingasmixtureareperformed.Usinga2m-longHC-PCFasgascell,thespectrumofacetyleneatn1+n3bandhasbeenmeasured,andtheP11-branchhasbeenselectedforthepurposeofsensing.Aminimumdetectivityof143partspermillionbyvolume(ppmv)forthesystemconfigurationisestimated.