基于三維特征臉模型的面部光照參數(shù)檢測技術研究.pdf

1、南開大學碩士學位論文基于三維特征臉模型的面部光照參數(shù)檢測技術研究姓名：張錚申請學位級別：碩士專業(yè)：計算機應用技術指導教師：白剛20070501AbstractFaeerecognitionwithvariantposeilluminalion趾dexpression(Pm)isaverychallengingproblemwhielac蛆notbeeasilysettledbytraditional2Dfacerecognitionte

2、elaniquesHavingconsideredthelimitationsof2DfacerecognitionWeaimatthechallengingproblem——facerecognitionwithvariantPIEtoexteadOUI“reseaxehproposingaframeworkof3DfacerecognitionsystemandfocusingOnthetwomostcriticalproblemo

3、f3Dfacerecognition：Firsthowtogoncratetheaverage3Dfacemodel，andsecondlyhowtecognizetheilluminationandposeofthetargetfacein鋤inputimageAccordingtotheresultofourexperimentOllrmethodisfeasibleandtheconclusion啪reachisverymeani

4、ngfulInthispaporatfirstWeU8ethemethodin【習based011quadrieeIrormelaiestosimplify0111“initial3Dfacemodel；ThenWeproposearesamplemethodbasedonkeyfeaturealignmenttodothepixeltopixelalignmentsamongthe3Dfaces，wlaiehisthekeyprobl

5、eminthepreproee88ingof3DfacemodelsComparedwithopticalflowalgorithm嗍wlaielaiscomplexandthe3Dresamplemethod【9】wlaiehneedsinteractionOUrmethodlassitsadvantagesand鯽伍cefor0111“workrequirements；LastlybyimposingdifferentPIEcond

6、itions011theaveragemodelandmatchingitwiththeinput2Dfaceilnage)werecognizethePIEintheinputfaceimageThematelaingisonlybasedonkeyfeaturespronetoPIEdifference∞it鋤nmatarelativehi曲speedKeyworaPIE，3Dfacerecognitionmeshsimplitie