简介：在这份报纸，许多三维(3D)数字胸模型基于临床的磁性的回声图象(MRI)被开发。一个混合轮廓察觉方法被用来创造轮廓，并且内部空间充满不同的胸纸巾，与相应于MRI象素紧张的指定间隔的各个。发达模型解剖地在乳房描述复杂织物结构和绝缘的性质。而且，他们与finite-difference-time-domain(FDTD)兼容格子房间。Convolutional完成式匹配的层(CPML)与FDTD一起被使用在模型外面模仿开的边界。在测试阶段，微波乳癌察觉模拟与改变X光线照相术的密度在四个模型被执行。然后，共焦的算法被利用重建肿瘤图象。成像结果证明肿瘤voxels能在每种情况中被认出，与在二个低密度盒子中的2公里地点错误和在二个高密度盒子中的7个mm8公里地点错误，证明导出MRI的模型能描绘病人乳房之间的单个差别。

简介：ＭＣＣＤＭＡ如同ＤＳＣＤＭＡ一样，也存在多用户干扰．ＭＬ准则是最佳的多用户检测准则，但其计算复杂度随用户数目增多而呈指数增长．神经网络利用其大规模并行的特点，可应用于此问题．本文论述了均场退火和混沌神经网络在此问题中的解决方案，并作了比较．

简介：许可证字符抽取的传统的方法不能满足识别精确性的要求，速度由录像显示车辆的察觉系统。因此，基于多尺度的边，察觉和一个字符分割算法基于Markov随机回答模型的一个牌照本地化方法被介绍。实验的结果证明方法基于阀值由差别操作员和特性分割基于边察觉与传统的本地化方法相对照产出更精确的许可证特性抽取。，精确性在差的状况下面在更好的照明下面从90%增加到94%，它增加超过5%。当二个改进算法被使用时，自动许可证识别的精确性和速度遇见系统“甚至在吵闹的情形或不平的照明下面的s要求。

简介：AgreatamountoffoodbornepathogenswereGram-positive(G?)bacteria,athreattopublichealth.Inthisstudy,consideringthebindingabilityofnisintowardsG-bacteriaandthestablefluorescentabilityofEGFPprotein,afluorescentnisin–EGFPproteinprobewasconstructedbyageneengineeringmethod.NisinandEGFPwereusedasthereceptorandfluorophore,respectively,todetectG?bacteria.ThenisinandegfpgenewereamplifiedseparatelyaccordingtothesequencepublishedinGenBankusinguniqueprimers.ThetwogeneswereclonedintoapET-28b(?)vectorresultinginapET-28b(?)–nisin–egfpvector.ThevectorwastransferredintoEscherichiacoli(E.coli)BL21(DE3)forexpression.Theexpressedproteinwasextracted,purifiedbyaNi–NTAcolumn,andthentestedbytheSDS-PAGEmethodtoconfirmitsmolecularweight.Listeriamonocytogenes(L.monocytogenes),Staphylococcusaureus(S.aureus),andMicrococcusluteus(M.luteus)wereusedastherepresentationsofG?bacteria.E.coliO157,representingthegram-negative(G-)bacteria,wasusedasanegativecontrol.Thebindingspecificityoftherecombinantproteinwasperformedontwotypesofbacteriaandthendetectedthroughfluorescentmicroscopy.Theresultsindicatedthatthenisin–EGFPprobecoulddetectG?bacteriaat108CFU/mL.

简介：Themarkingschememethodremovesthelowscoresofthecontractor'sattributesgivenbyexpertswhentheoverallscoreiscalculated,whichmayresultinthatacontractorwithsomelatentriskswillwintheproject.Inordertoremedytheabovedefectofthemarkingschememethod,anoutlierdetectionmodel,whichisonemissionofknowledgediscoveryindata,isestablishedonthebasisofthesumofsimilarcoefficients.Then,themodelisappliedtothehistoricalscoredataoftenderevaluationforci...

简介：Astructuredlow-rankmatrixrecoverymodelforRGBDsalientobjectdetectionisproposed.Firstly,theproblemisdescribedbyalow-rankmatrixrecovery,andthehierarchicalstructureofRGBimageisaddedtothesparsityterm.Secondly,thedepthinformationisfusedintothemodelbyaLaplacianregularizationtermtoensurethattheimageregionswhichsharesimilardepthvaluewillbeallocatedtosimilarsaliencyvalue.Thirdly,avariationofalternatingdirectionmethodisproposedtosolvetheproposedmodel.Finally,bothquantitativeandqualitativeexperimentalresultsonNLPR1000andNJU400showtheadvantageoftheproposedRGBDsalientobjectdetectionmodel.

简介：Pulse-pairs(PP)generatedbydistancemeasuringequipment(DME)causesevereinterferencetotheL-banddigitalaeronauticalcommunicationsystems.UnlikepreviousworksthatmainlyimprovedreceiverstructureformitigatingtheDMEinterference,thispaperproposesanovelandpracticaltimedomainPPmitigationmethodwithoutchangingtheexistingreceiverstructure.Ourmethodiscomposedofthreepartsandutilizesaspecificproperty,i.e.,arelativelyfixedwaveformofDMEPP.First,thecomplex-valuedwaveformofPPisautomaticallydetectedthroughaniterativeweightedaveragemethod.Then,onthebasisofthedetectedPPwaveform,eachPPisreconstructedaccordingtoitsamplitude,position,andinitialphase,whichareestimatedthroughasparserepresentationalgorithm.Finally,thereconstructedPPsaresubtractedfromthecontaminatedsignal.Numericalexperimentsshowthatcomparedwithrecentlypublishedmethodsthebiterrorrateobtainedbyourmethodisapproximately5dBbetter,whilethecomplexityismaintainedatthesamelevel.