基于细节融合的同卵双胞胎指纹识别框架(IJISA-V6-N1-10)

I.J.IntelligentSystemsandApplications,2014,01,84-101PublishedOnlineDecember2013inMECS()DOI:10.5815/ijisa.2014.01.10Copyright©2014MECSI.J.IntelligentSystemsandApplications,2014,01,84-101MinutiaeFusionBasedFrameworkforThumbprintIdentificationofIdenticalTwinsKamtaNathMishra1DepartmentofComputerScience&Engg.,B.I.T.Mesra,Ranchi(AllahabadCampus),IndiaE-mail:[e-mail]2DepartmentofComputerScience&Engg.,B.I.T.Mesra,Ranchi(AllahabadCampus),IndiaE-mail:[e-mail]3DepartmentofInformationTechnology,IndianInstituteofInformationTechnology,Allahabad,IndiaE-mail:[e-mail]4,AnkurSingh5DepartmentofComputerScience&Engg.,B.I.T.Mesra,Ranchi(AllahabadCampus),IndiaE-mail:[e-mail],[e-mail]−IdenticaltwinsidentificationisachallengingtaskbecausetheysharethesameDNAsequence.Thisresearchpaperpresentsminutiaecoordinatesandorientationanglesfusionbasedtechniqueforthumbprintidentificationofidenticaltwins.SixdifferentthumbprintimagesofidenticaltwinsweretakenatafixedtimeintervalusingH3T&Aterminal.Theminutiaecoordinatesandorientationanglesofthesethumbprintswerefusedtoformaunionset.Theunionsetvalueswerestoredinthesmartcardmemoryforfurtheridentification.Theminutiaecoordinatesandorientationanglesofathumbprintofthepersontobeidentifiedarecomputedandfusedtogetherforonlineidentification.Thefusedminutiaearecomparedwiththeminutiaeunionsetvaluesstoredinthesmartcardmemoryforidentityverification.Theproposedmethodwastestedonaselfgeneratedidenticaltwindatasetand50identicaltwinsofstandardFVC04andFVC06datasets.WeobservedinexperimentsthattheproposedmethodisaccuratelydifferentiatingtheidenticaltwinsofselfgeneratedandFVCdatasets.IndexTerms−MinutiaeCoordinates,OrientationAngles,MinutiaeUnionSet,IdenticalTwins,andThumbprintPatternsI.IntroductionSoftComputingisatoolwhichdealseffectivelywithuncertainty,vagueness,andimprecationanditprovidesalowcostsolution.Itplaysanindispensableroleincomputationalengineering.Softcomputingprovidesusanoptionforfindingoptimalsolutionswhichareunpredictable,uncertain,andpartiallytrue.Manysoftcomputingtechniques(ArtificialNeuralNetworks,GeneticAlgorithms,FuzzyLogicModels,andParticleSwarmTechniquesetc.)areusedinBiometricsystems.Wecandeterminethepartialtruevalueofbinarysetdependinguponcompletenessorincompletenessofresultsbyusingfuzzylogicmodelwhereasartificialneuralnetworksaredesignedforpatternreorganizationanddataclassification.Fuzzysetcanbeusedforstoringtheminutiaecoordinatesandorientationanglesintheformofroughsetortrainedsetwhichwillbeusefulforidentifyingidenticaltwins.Chenhasusedfuzzysetforstoringthelocaltrianglefeaturesofathumbprint[4,32].Neuronsarethebiologicalelementswhichareusedforinformationprocessinginthebrain.NeuralnetworkapproachwasfirstintroducedbyLeungin1990[20].Aneuralnetworkconsistsofanumberofheadsortailswhereheadsarethenodesandtailsarethelinks.Theneuralnetworkprovidesinterconnectionbetweennodes[21].Biometricsystemisamethodusedforidentifyingapersononthebasisofphysicalandbehavioralcharacteristics.Physiologicalfeaturesbasedbiometricsincludeirisimage,fingerprints,thumbprint,palmprint,face,fingerveins,handgeometry,DNAsequence,andpalates.Behavioralfeaturesbasedbiometricsincludesgaits(styleofwalkingandstyleoftalking),signature,bodylanguage,handwriting,andvoice[31].Twinsareclassifiedintwocategories:fraternaltwinsandidenticaltwins.WecaneasilyidentifyingtheMinutiaeFusionBasedFrameworkforThumbprintIdentificationofIdenticalTwins85Copyright©2014MECSI.J.IntelligentSystemsandApplications,2014,01,84-101fraternaltwinbecausetheydifferinfacestructure,DNAsequence,andgeneticcompositions.Identicaltwinsareproducedfromthesameegg.Therefore,theysharesameDNAsequenceandfacestructure[28,29].Identicaltwinscanbeidentifiedonthebasisoftheirthumbprints.Athumbprintimageincludesbifurcationpoints,ridges,leftloop,rightloop,arcandendpoints.Thesefeaturesofthumbprintsplayveryimportantroleinidenticaltwinsidentification[11].Thispaperclassifiesidenticaltwinsidentificationintwophases:Phase1:Trainthethumbprintdatasetsusingsoftcomputingtechniques.Phase2:Authenticatethepersonusingapredefinedprocess.Fig.1:RepresentsthesoftcomputingtechniquesusedbyresearchersforidentifyinganidenticaltwinThebreakagepointsandbifurcationpointsofathumbprintarecalledasminutiaepoints.Theorientationangleofaminutiapointiscalledasminutiaorientationangle.Thisresearchpaperusesminutiaecoordinatesandorientationanglesfusionbasedfuzzysettechniqueforidentificationofidenticaltwins.Inthenextsection(sectionII)wehavediscussedtheliteraturereviewrelatedtoidenticaltwinsidentification.ThesectionIIIdescribesamathematicalmodelofminutiaefuzzysetbasedthumbprintidentificationmethod.InsectionIVofthispaperwehaveproposedanovelminutiaefusionbasedunionsettechniqueforclassifyingidenticaltwins.ThesectionVdiscussestheexperimentalanalysisofproposedmethodonselfgeneratedandstandarddatasets.ThesectionVIdescribesconclusionsandfutureresearchwork.II.LiteratureReviewJungproposedarobustfingerprintidenti