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 > Object tracking with an adaptive color-based particle filter
侵权投诉

Object tracking with an adaptive color-based particle filter

Abstract. Color can provide an e#cient visual feature for tracking nonrigid objects in real-time. However, the color of an object can vary over time dependent on the illumination, the visual angle and the camera parameters. To handle these appearance chang

ObjectTrackingwithanAdaptive

Color-BasedParticleFilter

KatjaNummiaro1,EstherKoller-Meier2,andLucVanGool1,2

KatholiekeUniversiteitLeuven,ESAT/VISICS,Belgium

{knummiar,vangool}@esat.kuleuven.ac.be

SwissFederalInstituteofTechnology(ETH),D-ITET/BIWI,Switzerland

{ebmeier,vangool}@vision.ee.ethz.ch12

Abstract.Colorcanprovideane cientvisualfeaturefortrackingnon-rigidobjectsinreal-time.However,thecolorofanobjectcanvaryovertimedependentontheillumination,thevisualangleandthecameraparameters.Tohandletheseappearancechangesacolor-basedtargetmodelmustbeadaptedduringtemporallystableimageobservations.Thispaperpresentstheintegrationofcolordistributionsintoparticle lteringandshowshowthesedistributionscanbeadaptedovertime.Aparticle ltertracksseveralhypothesessimultaneouslyandweightsthemaccordingtotheirsimilaritytothetargetmodel.AssimilaritymeasurebetweentwocolordistributionsthepopularBhattacharyyacoe cientisapplied.Inordertoupdatethetargetmodeltoslowlyvaryingimageconditions,frameswheretheobjectisoccludedortoonoisymustbediscarded.

1Introduction

Objecttrackingisrequiredbymanyvisionapplications,butespeciallyinvideotechnology[2,5,7,12].Trackingmethodscanbedividedintotwomainclassesspeci edasbottom-uportop-downapproaches.Inabottom-upapproachtheimageisgenerallysegmentedintoobjectswhicharethenusedforthetracking.Forexampleblobdetection[12]canbeusedfortheobjectextraction.Incontrast,atop-downapproachgeneratesobjecthypothesesandtriestoverifythemusingtheimage.Typically,model-based[5,7]ortemplatematchingapproaches[2]comprisethisclass.Theproposedparticle lterwithcolor-basedimagefeaturesbelongstothetop-downapproachesastheimagecontentisonlyevaluatedatthehypotheticalobjectpositions.

Theideaofaparticle lter–toapplyarecursiveBayesian lterbasedonsamplesets–wasindependentlyproposedbyseveralresearchgroups[3,5,7,11].These ltersproviderobusttrackingframeworksastheyareneitherlimitedtolinearsystemsnorrequirethenoisetobeGaussian.Inthispaperwepresenttheintegrationofcolordistributionsintoparticle ltering,whichhastypicallyusededge-basedimagefeatures[4,5,7,13].Colorhistogramshavemanyadvantagesfortrackingnon-rigidobjectsastheyarerobusttopartialocclusion,arerotationandscaleinvariantandarecalculatede ciently.In[8]colorinformationhas

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