资源描述：
《基于形状的叶片图像检索-外文翻译 (2篇)》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、外文翻译毕业设计题目：植物叶片的形状特征提取及分类研究原文1：Shapebasedleafimageretrieval译文1：基于形状的叶片图像检索原文2：Featureextractionandautomaticrecognitionofplantleafusingartificialneuralnetwork译文2：利用人工神经网络实现植物叶片的特征提取与自动识别ShapebasedleafimageretrievalZ.Wang,Z.ChiandD.FengAbstract:Theauthorsprescntanefficienttwo
2、-stageapproachforleafimageretrievalbyusingsimpleshapefeaturesincludingcentroid-contourdistance(CCD)curve,eccentricityandanglecodehistogram(ACH).Inthcfirststage,theimagesthataredissimilarwiththequeryimagewillbefirstfiltcredoutbyusingecccutricicytoreducethesearchspace,andfine
3、retrievalwillfollowbyusingallthreesetsoffeaturesinthereducedsearchspaceinthesecondstage.DifferentfromeccentricityandACH,theCCDcurveisncitherscaling-invariantnorrotation-invariant.Therefore,normalisationisrequiredfortheCCDcurvetoachievescalinginvariance,andstartingpointlocat
4、ionisrequiredtoachieverotationinvariancewiththesimilaritymeasureofCCDcurves.Athinning-basedmethodisproposedtolocatestartingpointsofleafimagecontours,sothattheapproachusedismorecomputationallycfficient.Actually,themethodcanbencfitothershaperepresentationsthataresensitivetost
5、artingpointsbyreducingthematchingtimcinimagerecognitionandretrieval.Experimentalresultson1400leafimagesfrom140plantsshowthattheproposedapproachcanachieveabetterretrievalperformancethanboththecurvaturescalespace(CSS)methodandthemodifiedFourierdescriptor(MFD)method.Inaddition
6、,thetwostageapproachcanachieveaperformancecomparabletoanexhaustivesearch,butwithamuchreducedcomputationalcomplexity.1IntroductionPlantidentificationisaprocessresultingintheassignmentofeachindividualplanttoadescendingseriesofrelatedplantgroupsinternisoftheircommoncharacteris
7、tics.Suchataskisverydemandinginbiologyandagriculture,suchasforthediscoveryofnewspecies,plantresourcesurveysandplantspeciesdatabasemanagement.Sofar,thistime-consumingprocesshasmainlybeencarriedoutbybotanists.Asignificantimprovementcanbeexpectediftheplantidentificationcanbeca
8、rriedoutbyacomputer,automaticallyorsemautomatically,assistedbyimageprocessingandco
