Face Identification by Fitting a 3D Morphable Model

Face Identification by Fitting a 3D Morphable Model

ID:39221017

大小:561.17 KB

页数:15页

时间:2019-06-27

Face Identification by Fitting a 3D Morphable Model_第1页
Face Identification by Fitting a 3D Morphable Model_第2页
Face Identification by Fitting a 3D Morphable Model_第3页
Face Identification by Fitting a 3D Morphable Model_第4页
Face Identification by Fitting a 3D Morphable Model_第5页
资源描述:

《Face Identification by Fitting a 3D Morphable Model》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、FaceIdentificationbyFittinga3DMorphableModelusingLinearShapeandTextureErrorFunctionsSamiRomdhani,VolkerBlanz,andThomasVetterUniversityofFreiburg,Institut¨fur¨Informatik,Georges-Kohler¨-Allee52,79110Freiburg,Germany,fromdhani,volker,vetterg@informatik.uni-freiburg.deAbstractThispape

2、rpresentsanovelalgorithmaimingatanalysisandidenti-ficationoffacesviewedfromdifferentposesandilluminationconditions.Faceanalysisfromasingleimageisperformedbyrecoveringtheshapeandtexturesparametersofa3DMorphableModelinananalysis-by-synthesisfashion.Theshapeparametersarecomputedfromas

3、hapeerrorestimatedbyopticalflowandthetextureparametersareobtainedfromatextureerror.Thealgorithmuseslin-earequationstorecovertheshapeandtextureparametersirrespectiveofposeandlightingconditionsofthefaceimage.Identificationexperimentsarereportedonmorethan5000imagesfromthepubliclyavaila

4、bleCMU-PIEdatabasewhichincludesfacesviewedfrom13differentposesandunder22differentillumina-tions.Extensiveidentificationresultsareavailableonourwebpageforfuturecomparisonwithnovelalgorithms.1IntroductionImagesoffacesdonotdependonidentityonly,unfortunately.Theyarealsoinflu-encedbyillu

5、mination,bytheposefromwhichthefaceisviewedandbyitsexpression.Henceafaceanalysissystemisconfrontedwiththeperplexingtaskofseparatingdif-ferentsourcesofimagevariation.TherecentFERETtestsrevealedthatrecognitionperformancedropssignificantlywhenposeand/ordirectedlightarepresentintheinput

6、images[16].Solvingthesetwoproblemshasbecomeamajorresearchissue.Inthispaperwepresentanovelalgorithmabletoidentifyfacesinpresenceofcombinedposeandilluminationvariation.Acommonapproachforaddressingtheseproblemsistobuildanexplicitgenerativemodelofthevariationsoffaceimages.Then,analysi

7、sofanovelfaceimagerequiresthefittingofanimagegeneratedbythemodeltothenovelimage,therebycodingthenovelimageintermsofmodelparameters.Thisfittingproblemisinfactanoptimisa-tionproblem:Itsgoalistofindthemodelparameterswhichminimisethedifferencebetweentheimageproducedbythemodelandthenoveli

8、mage.Theminimisationcanbeperforme

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。