指纹识别系统原文.docx

指纹识别系统原文.docx

《指纹识别系统原文.docx》由会员分享，可在线阅读，更多相关《指纹识别系统原文.docx（18页珍藏版）》请在冰豆网上搜索。

指纹识别系统原文

ABSTRACT

WiththecontinuousdevelopmentofBiometricIdentificationTechnology,Peoplefoundthateachperson'sfingerprinthasuniquenessandinvariant.ThereforeFingerprintIdentificationTechnologygraduallydevelopedasanewidentityrecognitionmode，andwithitsgoodsafetyandreliability，ithasreplacedthetraditionalidentificationwaytrends.inbanks,homesetc.WehavetakenoneoftheBiometricmodalityi.e.fingerprintsforthispurposewhichisthemostreliablesecuritytoolforsecurityofthesystem.Fingerprintrecognitionhaslongbeenfavoredamongmanybiometricidentificationtechnologiesduetoitsuniquenessandpermanence.

Nowadays,fingerprintrecognitionisconsideredtobethebestchoiceformostapplicationsfromnetworksecuritysystemstocompactdevices,duetoitsaccuracy,speed,reliability,non-intrusiveinterfaces,andcost-effectiveness.WehavemadeafingerprintmatchingsystemusingMATLAB.AftercollectingthedatabaseweprovideasecuritytoavirtualprocessthathasbeendesignedinLABVIEWandmakeanefficientHMI.Thefingerprintrecognitionsystemmaysufferattacksatdifferentpointsduringtheauthenticationprocess.Somtimesthefingerprintsgives100%matchingbutitmaybefalsefingerprint.Themostcommonattacksoccurbytheuseoffakefingerprintduringthecaptureofimage.Thetransmissionchannelbetweenthefeatureextractorandmatchingmayalsobeinterceptedandthefingerprintfeaturemaybestoredforthelateruse.Themainproblemreliesonhowtodifferentiatealivefingerfromthatonemadeofsomesyntheticmaterial.InThisPaperwedisscusstheallsuchtypeofattacksaswellassomeotherintersetingandothertopicsrelatedtofingerprintslikethesemaybedamageorstolenbysomepersonandeffectofilllnesstothem.Thispapergivesanapprochtoremovesuchtypeofattacksbyusingavirtualsystem.Alongwiththisvirtualsystemsomeaspectsrelatedtosecurefingerprintisalsobeconsideredinthispaper.

Keywords：

Authentication,Biometrics,Fingerprintrecognition,HMI.

1.Introduction

Biometricsisatechnologythatuniquelyidentifiesapersonbasedonhisphysiologicalorbehavioralcharacteristics.Someofthebiometricswhichhaveexcellentpotentialforidentificationarethefingerprint,iris,retinalscan（Jainetal.,1997）.Fingerprintmatchingisoneoftheoldestandtrustedbiometricforidentification（LeeandGaensslen,1991）.Themanualclassificationsofthefingerprintsweretimeconsumingandpronetoerrors,henceitwasnotpossibletopromoteitonacommercialscale.ThefirstscientificpaperonfingerprintwasgivenbyEnglishmorphologistN.Grewin1684,inwhichhehadreportedhisstudyontheridge,furrowandporestructuresinthefingerprint.In1823,Purkinjedevelopedthefirstfingerprintclassificationschemeinwhichthefingerprintswereclassifiedintoninecategoriesaccordingtotheridgeconfiguration（LeeandGaensslen,1991）.SirF.Galtonafteranextensivestudyoffingerprintsintroducedtheminutiaefeaturesforthefingerprintclassificationintheyear1888.Theminutiaepointsarethelocaldiscontinuitiesintheridgeflowpattern.TheFBIinvestedalotofeffortsincetheearly1960sinthedevelopmentoftheAFIS（FederalBureauofInvestigation,1984）.EversincetheAutomaticFingerprintIdentificationSystem（AFIS）wasdevelopedin1991,rapidprogresshasbeenmadeinimprovingtherecognitionrates.Thisdevelopmenthasgreatlyimprovedthescopeofusingfingerprintforidentificationpurposesandinreducingthecostofhiringandtraininghumanexpertsforfingerprintmatching.Thecurrentworksinthisfieldfocusonmainlyreducingthecomputationtimeforfeatureextractionandmatchingandimprovingtheaccuracyoftheresults.Today,withthecrimeratesincreasingeverydaythereisanurgentneedforasystemwhichissafeandfast.

Securityisanevergrowingissueintoday’sworld.Itiscommontosafeguardthecomputers,offices,etc.Previously,passwordsandpinswerewidelyusedforprotection,butlateron,itwasfoundoutthatthesewerepronetoabuse.Thus,personalbiometricrecognitionhascometothestageandcurrentlyfingerprintandirisrecognitionarethemostwidelyusedmethods.Thosearealsousedincriminalcases,whichmakeserrorminimizationextremelyimportant.Therestillexiststheneedforbetterhardwareandsoftware.

Inthispaper,anovelandefficientalgorithmwhichrecognizesagivenfingerprintwithhighaccuracyispresented.Thesimulationresultsaregiven.Comparisonwiththepreviouslydesignedalgorithmsismadebasedonensembleaveraging.HMFAhasbeenfoundtoyieldmuchbetterresults。

2.INTRODUCTION

2.1Basicsonfingerprint

Fingerprintsarethetinyridges,whorlsandvalleypatternsonthetipofeachfinger.Theyformfrompressureonababy'stiny,developingfingersinthewomb.Notwopeoplehavebeenfoundtohavethesamefingerprints--theyaretotallyunique.There'saonein64billionchancethatyourfingerprintwillmatchupexactlywithsomeoneelse's.FingerprintsareevenmoreuniquethanDNA,thegeneticmaterialineachofourcells.AlthoughidenticaltwinscansharethesameDNA--oratleastmostofit--theycan'thavethesamefingerprints.Fingerprintingisoneformofbiometrics,asciencethatusespeople'sphysicalcharacteristicstoidentifythem.Fingerprintsareidealforthispurposebecausethey'reinexpensivetocollectandanalyze,andtheyneverchange,evenaspeopleage.

Fingerprintsaremadeofanarrangementofridges,calledfrictionridges.Eachridgecontainspores,whichareattachedtosweatglandsundertheskin.Youleavefingerprintsonglasses,tablesandjustaboutanythingelseyoutouchbecauseofthissweat.Thefingerprintsurfaceismadeupofasystemofridgesandvalleysthatserveasfrictionsurfacewhenwearegrippingtheobjects.Thesurfaceexhibitsveryrichstructuralinformationwhenexaminedasanimage.Thefingerprintimagescanberepresentedbybothglobalaswellaslocalfeatures.Theglobalfeaturesincludetheridgeorientation,ridgespacingandsingularpointssuchascoreanddelta.Thesingularpointsareveryusefulfromtheclassificationperspective.However,verificationusuallyreliesexclusivelyonminutiaefeatures.Minutiaearelocalfeaturesmarkedbyridgediscontinuities.Thereareabout18distincttypesofminutiaefeaturesthatincluderidgeendings,bifurcations,crossoversandislands.Amongthese,ridgeendingsandbifurcationarethecommonlyusedfeatures.Therearetwodifferenttypesofprints:

Visibleprintsaremadeonatypeofsurfacethatcreatesanimpression,likeblood,dirtorclay.

Latentprintsaremadewhensweat,oilandothersubstancesontheskinreproducetheridgestructureofthefingerprintsonaglass,murderweaponoranyothersurfacetheperpetratorhastouched.Theseprintscan'tbeseenwiththenakedeye,buttheycanbemadevisibleusingdarkpowder,lasersorotherlightsources.Policeofficerscan"lift"theseprintswithtapeortakespecialphotographsofthem.

Basedontwentyyearsofcontinuingtechnologydevelopmentandsuccessfuloperationalexperience,fingerprintbiometricsremaintheonlyscientificallyandoperationallyprovenbiometrictechnology,theonlybiometrictechnologywithamatureinfrastructureofsupportingtechnologies,andtheonlytechnologysupportedbynationalandinternationalstandards.Technologyadvancesmadeoverthepastfewyearshaveresultedinoff-the-shelffingerprintIDsystemswiththecapabilitytocompleteduplicateregistrationsearchesofverylargedatabasesinseconds.Fingerprintshaveseveraladvantagesoverotherbiometrics,suchasthefollowing:

1.Highuniversality:

Alargemajorityofthehumanpopulationhaslegiblefingerprintsandcanthereforebeeasilyauthenticated.Thisexceedstheextentofthepopulationwhopossesspassports,IDcardsoranyotherformoftokens.

2.Highdistinctiveness:

EvenidenticaltwinswhosharethesameDNAhavebeenshowntohavedifferentfingerprints,sincetheridgestructureonthefingerisnotencodedinthegenesofanindividual.Thus,fingerprintsrepresentastrongerauthenticationmechanismthanDNA.

3.Highpermanence:

Theridgepatternsonthesurfaceofthefingerareformedinthewombandremaininvariantuntildeathexceptinthecaseofsevereburnsordeepphysicalinjuries.

4.Highperformance:

FingerprintsremainoneofthemostaccuratebiometricmodalitiesavailabletodatewithjointlyoptimalFAR（falseacceptrate）andFRR（falserejectrate）.

5.Wideacceptability:

Whileaminorityoftheuserpopulationisreluctanttogivetheirfingerprintsduetotheassociationwithcriminalandforensicfingerprintdatabases,itisbyfarthemostwidelyusedmodalityforbiometricauthentication.Althoughaswedisscussedfingerprintismostreliablebiomericsforauthenticationbutbecauseofhighcrimnalactivitiestheremaybeattackonthesefingerprints.Thefingerprintrecognitionsystemmaysufferattacksatdifferentpointsduringtheauthenticationprocess.Sectionandsub-sectionheadingsandttributes.Themostcommonattacksoccurbytheuseoffakefingerprintduringthecaptureofimage.Afakefingerprintarebuildfromlatentfingerprintleftattoucheditemssuchasglasses,doorknobs,glossypaper,etc.Usingthisfingerprintarebuildthree-dimensionalmoldsofrubbermembrane,glue,orgelatin.Duringthetransmissionoftheimagetothefeatureextractormayoccurinterceptionofthechannel,andconsequently,thefingerprintimagemaybestolenandlater,usedforfakefingerprintconstructionorfordirectlyaccesstofeatureextractorbybypassingthescanner.Thefeatureextractormaybesubstitute