基于膚色和支持向量機(jī)的人臉檢測技術(shù)研究.pdf

1、武漢理工大學(xué)碩士學(xué)位論文基于膚色和支持向量機(jī)的人臉檢測技術(shù)研究姓名：穆麗萍申請學(xué)位級別：碩士專業(yè)：控制理論與控制工程指導(dǎo)教師：陳定方20070501武漢理工大學(xué)碩士學(xué)位論文AbstractFacedetectioniSatechniquetodetecthumanfacesinapictureandusecertainstrategytolabelitslocationsizeandpositionItisthefirstonetoso

2、lveinmanyhumanfacesinformationprocessingdomainItisessentialtealizetheautomaticpersonfacerecognitionsystemItismorethan20yearsabouttheresearchforpCl∞nfacedetection；sofarbecauseofthecomplexityofpersonfacedetectionitselfther

3、ehasn’tbeenamethodCansolveittotallyPeoplestartedtothinkaboutitfromdifferentwaysandconsideringthefusionofmanydifferentmethodsallthesehadobtainedquitewellresultsAgreatamountofliteratures，surveysandresearchpapersconcerningu

4、ptodatetechniquesoffacedetectionandfacerecognitionarcreadandanalyzedSomehotissuesaboutfacedetectionarediscussedAfacedetectionsystemaboutcolorinlagesisbuiltup11他researchworkofthispapermainlyincludesthefollowingfouraspects

5、：(1)IthasrcalizgdaskinregiondivisionalgorithmbasedOntheYCbCrcolorspaceInviewofthetraditionalskincolormodel，Ihavemadesomeimprovements；ithasenhancedthedetectionspeedefficientlywithoutaffectingtheresultwhichisadvantageousto

6、therealizationofrealtimepersonfacedetectionsystem(2)IhaveusedthetraditionalvalvevaluedivisionthoughtinskincolorregionprocessingTothesameimage，onekindoftimesharingdynamicvalvevaluedivisionalgorithmhasbeenusedItcaneffectiv

7、elyseparatethedifferentskincolorregionandtheskincolorregionfIomthebackgroundinthebasisofdetectouteveryskinregion；Tothedifferentimages，onekindofoverallsituationdynamicvalvevaluedivisionalgorithmhasbeenused。ThismethodcaIla

8、dapttoallkindsofdifferentsizesofpersonfacesinthesameordifferentimagesverywell(3)Ithasrealizedthefinedetcc時onandlocalizationofpersonfacebasedonthesupportvectormachine(SVM)method，眥smethodisusedbasedontheskincolordivisionpr