文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 自然科学 > Efcient Head Pose Estimation with Gabor Wavelet Networks
侵权投诉

Efcient Head Pose Estimation with Gabor Wavelet Networks

Efcient Head Pose Estimation with Gabor Wavelet Networks

Ef cientHeadPoseEstimationwithGabor

WaveletNetworks

VolkerKrueger,SvenBrunsandGeraldSommer

ComputerScienceInstitute,Christian-AlbrechtsUniversityKiel

Preußerstr.1-9,24105Kiel,Germany

Tel:++49-431-560496,FAX:++49-431-560481

email:vok@ks.informatik.uni-kiel.de

Abstract

Inthisarticlewewanttointroduce rsttheGaborwaveletnetworkasamodelbasedapproachforaneffectiveandef cientobjectrepresentation.TheGa-borwaveletnetworkhasseveraladvantagessuchasinvariancetosomede-greewithrespecttotranslation,rotationanddilation.Furthermore,theuseofGabor ltersensuredthatgeometricalandtexturalobjectfeaturesareen-coded.ThefeasibilityoftheGabor ltersasamodelforlocalobjectfea-turesensuresaconsiderabledatareductionwhileatthesametimeallowinganydesiredprecisionoftheobjectrepresentationrangingfromasparsetoaphoto-realisticrepresentation.InthesecondpartofthepaperwewillpresentanapproachfortheestimationofaheadposethatisbasedontheGaborwaveletnetworks.

1Introduction

Recently,model-basedapproachesfortherecognitionandtheinterpretationofimagesofvariableobjects,likethebunchgraphapproach,PCA,eigenfacesandactiveappearancemodels,havereceivedconsiderableinterest[23;14;4;6].Theseapproachesachievegoodresultsbecausesolutionsareconstrainedtobevalidinstancesofamodel.Intheseapproaches,theterm“model-based”isunderstoodinthesensethatasetoftrainingob-jectsisgivenintheformofgrayvaluepixelimageswhilethemodel“learns”thevariancesofthegrayvalues(PCA,eigenfaces)or,respectively,theGabor lterresponses(bunchgraph).Withthis,modelknowledgeisgivenbythevariancesofpixelgrayvalues,whichmeansthattheactualknowledgerepresentationisgivenonapixelbasis,thatisindepen-dentfromtheobjectsthemselves.

InthisworkwewanttointroduceanovelapproachforobjectrepresentationthatisbasedonGaborWaveletNetworks.GaborWaveletNetworks(GWN)arecombiningtheadvantagesofRBFnetworkswiththeadvantagesofGaborwavelets:GWNsrepresentanobjectasalinearcombinationofGaborwaveletswheretheparametersofeachoftheGaborfunctions(suchasorientationandpositionandscale)areoptimizedtore ecttheparticularlocalimagestructure.Gaborwaveletnetworkshaveseveraladvantages:

1.Bytheirverynature,Gaborwaveletnetworksareinvarianttosomedegreetoaf nedeformationsandhomogeneousilluminationchanges,

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email