Fast template matching using bounded partial correlation

Fast template matching using bounded partial correlation

ID:39369076

大小:264.93 KB

页数:9页

时间:2019-07-01

Fast template matching using bounded partial correlation_第1页
Fast template matching using bounded partial correlation_第2页
Fast template matching using bounded partial correlation_第3页
Fast template matching using bounded partial correlation_第4页
Fast template matching using bounded partial correlation_第5页
资源描述:

《Fast template matching using bounded partial correlation》由会员上传分享,免费在线阅读,更多相关内容在应用文档-天天文库

1、MaschineVisionandApplications(2003)13:213–221MachineVisionandApplicationscSpringer-Verlag2003FasttemplatematchingusingboundedpartialcorrelationLuigiDiStefano,StefanoMattocciaDEIS-ARCES,UniversityofBologna,VialeRisorgimento2,40136Bologna,Italy(e-mail:{ldistefano,smattoccia}@deis

2、.unibo.it)Received:2November2000/Accepted:25July2001Abstract.Thispaperdescribesanovel,fasttemplate-Sincewithlarge-sizeimagesandtemplatesthematch-matchingtechnique,referredtoasboundedpartialcorrelationingprocesscanbecomputationallyveryexpensive,numer-(BPC),basedonthenormalisedcro

3、ss-correlation(NCC)func-oustechniquesaimedatspeedingupthebasicapproachhavetion.Thetechniqueconsistsincheckingateachsearchposi-beendevised(see[4]foraconcisereview).Amonggeneraltionasuitableeliminationconditionrelyingontheevaluationtechniques(i.e.applicablewithbothdistorsionandcor

4、relationofanupper-boundfortheNCCfunction.Thecheckallowsmeasures),themajoronesare(a)theuseofmulti-resolutionforrapidlyskippingthepositionsthatcannotprovideabetterschemes[5](i.e.locatingacoarse-resolutiontemplateintodegreeofmatchwithrespecttothecurrentbest-matchingone.thecoarse-re

5、solutionimageandthenrefiningthesearchatTheupper-boundingfunctionincorporatespartialinformationthehigherresolutionlevels),(b)sub-samplingtheimageandfromtheactualcross-correlationfunctionandcanbecalcu-thetemplate,(c)two-stagematching[7](i.e.matchingasub-latedveryefficientlyusingarec

6、ursivescheme.Weshowalsotemplatefirst,andthenthewholetemplateonlyatgoodcan-asimpleimprovementtothebasicBPCformulationthatpro-didatepositions).However,techniques(a),(b)and(c)implyavidesadditionalcomputationalbenefitsandrendersthetech-non-exhaustivesearchprocesssincetheydonotcomparet

7、heniquemorerobustwithrespecttotheparameterschoice.full-resolutionimagewiththefull-resolutiontemplateatev-erysearchposition.Asaresult,thereisnoguaranteethattheKeywords:Templatematching–Patternmatching–Nor-algorithmfindstheglobaldistorsionminimum(orcorrelationmalisedcross-correlati

8、on–SEA–PDEmaximum).Ontheotherhand,inthespecificc

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。