图像匹配是NP完全问题外文文献翻译.doc

图像匹配是NP完全问题外文文献翻译.doc

ID:59217957

大小:843.50 KB

页数:28页

时间:2020-09-09

图像匹配是NP完全问题外文文献翻译.doc_第1页
图像匹配是NP完全问题外文文献翻译.doc_第2页
图像匹配是NP完全问题外文文献翻译.doc_第3页
图像匹配是NP完全问题外文文献翻译.doc_第4页
图像匹配是NP完全问题外文文献翻译.doc_第5页
资源描述:

《图像匹配是NP完全问题外文文献翻译.doc》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库

1、ElasticimagematchingAbstractOnefundamentalprobleminimagerecognitionistoestablishtheresemblanceoftwoimages.Thiscanbedonebysearchingthebestpixeltopixelmappingtakingintoaccountmonotonicityandcontinuityconstraints.WeshowthatthisproblemisNP-completebyreductionfrom3-SAT,thusgivingevidencethattheknow

2、nexponentialtimealgorithmsarejustified,butapproximationalgorithmsorsimplificationsarenecessary.Keywords:Elasticimagematching;Two-dimensionalwarping;NP-completeness1.IntroductionInimagerecognition,acommonproblemistomatchtwogivenimages,e.g.whencomparinganobservedimagetogivenreferences.Inthatpro-

3、cess,elasticimagematching,two-dimensional(2D-)warping(UchidaandSakoe,1998)orsimilartypesofinvariantmethods(Keysersetal.,2000)canbeused.Forthispurpose,wecandefinecostfunctionsdependingonthedistortionintroducedinthematchingandsearchforthebestmatchingwithrespecttoagivencostfunction.Inthispaper,we

4、showthatitisanalgorithmicallyhardproblemtodecidewhetheramatchingbetweentwoimagesexistswithcostsbelowagiventhreshold.WeshowthattheproblemimagematchingisNP-completebymeansofareductionfrom3-SAT,whichisacommonmethodofdemonstratingaproblemtobeintrinsicallyhard(GareyandJohnson,1979).Thisresultshowst

5、heinherentcomputationaldifficultiesinthistypeofimagecomparison,whileinterestinglythesameproblemissolvablefor1Dsequencesinpolynomialtime,e.g.thedynamictimewarpingprobleminspeechrecognition(seee.g.Neyetal.,1992).Thishasthefollowingimplications:researcherswhoareinterestedinanexactsolutiontothispr

6、oblemcannothopetofindapolynomialtimealgorithm,unlessP=NP.Furthermore,onecanconcludethatexponentialtimealgorithmsaspresentedandextendedbyUchidaandSakoe(1998,1999a,b,2000a,b)maybejustifiedforsomeimagematchingapplications.Ontheotherhandthisshowsthatthoseinterestedinfasteralgorithms––e.g.forpatter

7、nrecognitionpurposes––arerightinsearchingforsub-optimalsolutions.Onemethodtodothisistherestrictiontolocaloptimizationsorlinearapproximationsofglobaltransformationsaspresentedin(Keysersetal.,2000).Anotherpossibilityistouseheuristicapproa

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。