简介:为了更方便地并且更快速理解超声图象的内容,在这份报纸,我们建议一个coarse-to-fine超声图象captioning整体模型,它能自动地产生由相关n克构成在超声图象描述疾病信息的注解文章。首先,在超声图象的机关被粗糙的分类模型检测。第二,超声图象被相应有细密纹理的分类模型根据机关标签编码。最后,我们输入编码向量到产生模型自动地产生的语言产生模型,和语言在超声描述疾病信息的注解文本想象。在我们的实验,编码模型能在超声图象识别获得高精确性率。并且语言产生模型能自动地产生高质量的注解文本。在实际应用,coarse-to-fine超声图象captioning整体模型能帮助病人和医生获得很好超声图象的内容理解。
简介:我们调查用操作的一个特殊的类列出联合的问题,前缀移动。联合作为0和1的位串被代表,并且前缀移动是旋转的操作由一个位置到的一点绳的某前缀左或正确。我们把一个否定答案给F问的一个开的问题。Ruskey和威廉斯·阿(由前缀移动产生联合,在Proc。第11年度国际计算和组合数学会议2005,LNCS3595,Springer,2005,pp.570576),那是我们是否能由仅仅在位串上使用三很基本的前缀移动产生联合,它是开始的二位和由在任何一个方向的一个位置的全部位串的旋转的调换(即,使用排列σ2,σn和σn?1到位串的索引)。电子增补材料这篇文章(doi:10.1007/s11390-007-9094-7)的联机版本包含增补材料,它对授权用户可得到。
简介:Kaup-Newell层次能从是宽松的integrable的所谓的产生方程被导出,这被显示出。积极;在层次的否定流动同时被导出。产生方程;这些流动的相互的交换性使我们能构造新宽松的integrable方程。
简介:ThispaperpresentsthedataonoperationreliabilityindicesandrelevantanalysestowardChina'sconventionalpowergeneratingunitsin2009.Theunitsbroughtintothestatisticalanalysisinclude100-MWorabovethermalgeneratingunits,40-MWorabovehydrogeneratingunits,andallnucleargeneratingunits.Thereliabilityindicesembodiedincludeutilizationhours,timesandhoursofscheduledoutages,timesandhoursofunscheduledoutages,equivalentforcedoutagerateandequivalentavailabilityfactor.
简介:WeproposethevariationaldescriptionofgeneratingfunctionapproachoffirstkindforHamil-tonianODEs,andextendtheapproachtothesemi-linearwaveequations.Inthisway,wecanconstructanyfiniteorderaccuracyscheme,andshowthattheresultingnumericalschemeismultisymplectic.Atlast,wepresentsomenumericalexperimentsbyusingderivednewscheme.
简介:Hypemecrospenniaisthecommondiseasefrequentlyseeninandriatry.It'soneofthemainreasonsthatcausethemasculinesterility.DuringMarch1997December2003.Theauthorapplieddecoctionfortonifyingthekidneyandgeneratingthespermintreatmentofthisdiseaseandreceivedrathergoodefficacy.Thecompletesummaryof136caseswasreportedasfollowed.
简介:Baosteel's60000m3/hairseparationunit(ASU)isthelargestoxygengeneratingsysteminChina.Theoperationalcostofsuchagiantsystemisveryhigh.Howtoreducetheoperationalcostisacriticalissue.Thispaperdiscussesthesystem'scharacteristics,thecurrentoperationalstatusandthedifficultiesinreducingthecost,andanalyzesrelevantindicators,suchasthetechnicalandeconomicalindicatorsofindividualunitsandsystemsaswellastheindicatorsconcerningthecosts.Therelationshipbetweenthecostandeacheconomicalindicatorandmeasurestooptimizeaneconomicaloperationoftheoxygengeneratingsystemarealsodiscussedinthispaper.
简介:从截断的multivariate正常分布(TMVND)的采样组成在适合统计的许多和econometric的计算模块建模的核心。我们建议二个有效方法,一个反复的数据扩大(DA)算法和非反复的逆Bayes公式(IBF)刺绣花样,模仿TMVND并且与线性不平等限制概括他们到multivariate正常分布。由创造贝叶斯的不完全数据的结构,DA算法的以后的步直接产生随机的向量与单个元素对比绘画绘画,产生明显的计算优点并且与象S加,MATLAB和高斯那样的普通统计软件包裹的容易的编码。而且,DA为实现一个快他们算法识别TMVND的模式提供准备好了的结构,它在抑制参数问题的统计推理有许多潜在的应用程序。另外,作为中间的结果利用这个模式,IBF采样提供一种新奇选择给吉布斯采样并且由于在TMVND的部件之间的高关联与集中和可能的慢集中消除问题。DA算法与抑制参数被用于一个线性回归模型并且与一个出版数据集合被说明。数字比较证明建议DA算法和IBF刺绣花样比吉布斯刺绣花样和接受拒绝算法更有效。
简介:
简介:Inthispaper,twoalgorithmsarepresentedforgeneratingtwocodescanlistsofanN-dimensionalHilbertcell,andaformalproofofthebackwardencodingalgorithmisgiven.Onthebasisoftheself-simi-laritypropertiesofaHilbertcurve,thispapergivesanovelalgorithmforgeneratingastaticevolvementruletablethroughanalyzingaHilbertcell.Bylookingupthestaticevolvementruletable,theN-dimensionalHilbertmappingsareefficientlyimplemented.
简介:Majorsocietalproblemsaffectthesocialstability.Itisnecessarytounderstandthepublicopiniontowardthoseissuestoavoidsocialconflicts.Nowadaysthesocialmediabecomethemajorplatformtotrackwhatthepublicisconcernedaboutandwhichmaybeofthesocietalrisk.However,itisverytoughtocapturethepublicattentioninshorttimeduetohugeflowofuser-generatedcontents.Inthispaper,weapproachthisproblembyexpandingthemethodofgeneratingstorylinewiththeresultdisplayedbyamulti-viewgraph.Onereal-worldexampleisillustratedandevaluationisgiventoshowtheeffectivenessoftheproposedmethod.
简介:Griddispersioningeneratingfinite-differencessyntheticseismogramsAbdolrahimJavaherian(InstituteofGeophysics,theUniversityofTeh...
简介:AnApproximateAlgorithmofGeneratingVariateswithArbitraryContinuousStatisticalDistributions¥ZhangZhengjun;YangZiqiang;ZhangChun...
简介:Selectionoftestcasesplaysakeyroleinimprovingtestingefficiency.Black-boxtestingisanimportantwayoftesting,anditsvalidityliesontheselectionoftestcasesinsomesense.Areasonableandeffectivemethodabouttheselectionandgenerationoftestcasesisurgentlyneeded.Thisletterfirstintroducessomeusualmethodsonblack-boxtestcasegeneration,thenproposesanewalgorithmbasedoninterfaceparametersanddiscussesitsproperties,finallyshowstheeffectivenessofthealgorithm.
简介:Acontourtreeisagoodgraphicaltoolforrepresentingthespatialrelationsofcontourlinesandhasmanyapplicationsinmapgeneralization,mapannotation,terrainanalysis,etc.AnewmethodforgeneratingcontourtreesbyintroducingaVoronoi-basedinterioradjacencyconceptisproposedinthispaper.Theimmediateinterioradjacencysetisemployedtoidentifyallofthechildrencontoursofeachcontourwithoutcontourelevations.Ithasadvantagesoverexistingmethodssuchasthegeometricmethodandtheregiongrowingbasedmethod.