基于gabor小波和非線性算法的人臉識別系統(tǒng)

1、重慶大學碩士學位論文基于Gab小波和非線性算法的人臉識別系統(tǒng)姓名：樊超申請學位級別：碩士專業(yè)：儀器科學與技術(shù)指導教師：李見為20090527重慶大學碩士學位論文英文摘要ABSTRACTFacerecognitionisallidentificationtechnologywhichanalyzesfaceimagewithcomputerandthengettingeffectiverecognitioninformationWitht

2、hedevelopmentofsomeapplicationssuch鶴electroniccoln／1lgrce，facerecognitionhasbecomeoneofthemostpotentialbiometricsauthenticationmethodsitCanbeappliedinmanyaspects，such嬲criminalrecognitionofpolicesystem，monitorsystemofbank

3、andcustom，automaticdoormansystemandfamilyentertainment，etcThefacerecognitionsystemincludestwoparts：facedetectionandfacerecognitionFacedetectionistheInstandkeystepoffacerecognitionsystem，theprecisionoffacedetectionalgorit

4、hmdirectlyaffectsthefollowingimage’SdisposalandanalysisstepsThispaperadoptsthefacedetectionalgorithmbasedonAdaBoostwhichiseffectivelyforprophaseworkofthefacerecognitionsystemTheimportanceoffacerecognitionisfeatureextract

5、ionInthispaperweanalysismanyoffacerecognitionalgorithms，andthenaccordingtotheresultofexperiments，wepresentanewfacerecognitionalgorithmbasedonGaoborwaveletandSupervisedLocalLinearEmbedding，thisalgorithm’Seffectivenesshasb

6、eenverifiedbyexperimentsThispaper’Sprimaryworkasfollowing：1)AnalysisandrealizethefacedetectionalgorithmbasedonAdaBoostAndthentheexperimentsonthepicturesfrominternetshowtheadvantageofthisflgofithm2)Introducethedefinitiono

7、fGaborTransformandthemethodoffeatureextractionbyusingGaborwavelettransforillTheGaborwavelettramformhastwoadvantages：Firstlyitcondescribethereceptivefieldofsimplecellsinthevisualcortexofhumancerebracorrectly；SecondlyGabor

8、waveletsshowsthejointtime—frequencypropertyofsignalanalysisandovercomestheshortcomingsoftraditionalFourierTransformwhilefailstopresentanytimediscriminationabilityinfrequencydomainAnalysiscoefficientsvalueof2DGaborwavelet

9、anditsapplicationinfacerecognitionemphatically3)Introduceandanalysismanyclassicfacerecognitionalgorithms，suchasPCA、LDA、KPCA、KLDA、LLE，etcAccordingtotheexperimentsconductontheMatlabplatform，sortingouttheadvantageandremaini

10、ngproblemofeachalgorithmInordertosolvetheproblemofthesampleinformationnotbeingfullyusedanddeficienciesofotheralgorithmwhichbepresentedbefore，afacerecognitionalgorithmbaseonGaborwaveletandSLLE(SupervisedLocallyLinearEmbed