文档介绍：AbstractRecentyears,putertechnology,theFaceRecognition(FR),prisesaseriesofrelevanttechnologiesmakinguptheFRSystem,includingFaceImageAcquisition,FaceLocation,FRPre-processing,IdentityCheckandIdentityFinder,etc;-,allkindsofFRdeviceshavebeenwidelyusedinlotsofpracticalfields,suchastheEnterpriseEntranceGuardSystem,StudentTimeandAttendanceSystem,PoliceSupervisingSystemandIntelligenceCamera,,in2008,urredinOlympicGamesandplayedanimportantroleinsafeguardingtheopeningandclosingceremoniesofBeijingOlympics’GamesandParalympics’,theFRtechnologystillhasalotofproblems,whichcanbereducedtotwoclasses:inter-classdifferenceandintra-,theotheristhatthefacevariationsfromoneperson,suchasthevariationsresultedfromdifferentconditionsoffaceexpressions,visualangles,lightening,,theSIFTalgorithm,,thispaper,basedonthewideapplicationofthelocalfeatureoperatorsinFRfield,seeingthatSIFToperatorcanworkverywellontherecognitionandmatchingofrigidobjects,petenciestoFRsystemandsolvethehotproblemsexistedatthepresentstage,suchas:distortion,occlusionandillumination,:binedwithapatternmatchingstrategybasedonK-meansclusteringforFR,furthermore,binedthevectorspacesimilaritywiththekeypointsmatchingsimilaritytocalculatetheglobalsimilaritymatchingforfaceimages,andmakeuseofthesquareofsimilarityandtheweight-’resultshav