简介：等轴的设计(IsoProjection)是一个线性维数减小方法，它明确地考虑歧管在数据嵌入的结构。然而，IsoProjection是非直角的，它使它极其敏感到减少的空间的尺寸并且对估计困难内在的维数。non-orthogonality也弄歪在数据嵌入的公制的结构。这份报纸建议一个新方法叫了直角的等轴的设计(O-IsoProjection)，它分享象IsoProjection的一样的线性特性并且克服IsoProjection的公制的失真问题。类似于IsoProjection，O-IsoProjection第一构造能思考的一张毗邻图歧管在数据嵌入的结构和在样品之间的类关系脸空格指，然后由保存如此的图结构获得设计。与IsoProjection不同，O-IsoProjection要求直角的基础向量，和直角的基础向量能由反复的方法是计算的。ORL和耶鲁数据库上的试验性的结果证明O-IsoProjection让更好的识别比Eigenface，Fisherface和IsoProjection为脸识别评价。

简介：极端宽带(UWB)要求的高速度的高分辨率的analog-to-digital(A/D)变换信号处理是一个很挑战性的问题。这份报纸为UWB信号获得建议一个平行随机的设计方法。建议方法能完成高采样率，高决定和硬件实现的技术可行性。在建议方法，一个模拟UWB信号在一套随机的符号函数上被投射。然后，低率的高分辨率的analog-to-digital变流器(模数转换器)被用来取样设计系数。信号能被简单线性计算与采样矩阵重建，没有遵循优化算法和优先的知识。在另外的方面，不同其它走近那需要在极其高的频率利用精确时间移动，或设计一个混合过滤器银行，或与优先的知识为信号产生特定的基础功能或工作，建议方法是一条通用采样途径并且对容易适用。到噪音比率(SNR)和假免费的动态范围(SFDR)的信号的模拟结果为UWB信号获得验证建议方法的效率。

简介：Previously,anovelclassifiercalledKernel-basedNonlinearDiscriminator(KND)wasproposedtodiscriminateapatternclassfromotherclassesbyminimizingmeaneffectofthelatter.Toconsidertheeffectofthetargetclass,thispaperintroducesanobliqueprojectionalgorithmtodeterminethecoefficientsofaKNDsothatitisextendedtoanewversioncalledextendedKND(eKND).IneKNDconstruction,thedesiredoutputvectorofthetargetclassisobliquelyprojectedontotherelevantsubspacealongthesubspacerelatedtootherclasses.Inaddition,asimpletechniqueisproposedtocalculatetheassociatedobliqueprojectionoperator.ExperimentalresultsonhandwrittendigitrecognitionshowthatthealgorithmperformesbetterthanaKNDclassifierandsomeothercommonlyusedclassifiers.

简介：Theperformanceoftheclassicalclusteringalgorithmisnotalwayssatisfiedwiththehigh-dimensionaldatasets,whichmakeclusteringmethodlimitedinmanyapplication.Tosolvethisproblem,clusteringmethodwithProjectionPursuitdimensionreductionbasedonImmuneClonalSelectionAlgorithm(ICSA-PP)isproposedinthispaper.ProjectionpursuitstrategycanmaintainconsistentEuclideandistancesbetweenpointsinthelow-dimensionalembeddingswheretheICSAisusedtosearchoptimizingprojectiondirection.Theproposedalgorithmcanconvergequicklywithlessiterationtoreducedimensionofsomehigh-dimensionaldatasets,andinwhichspace,K-meanclusteringalgorithmisusedtopartitionthereduceddata.TheexperimentresultsonUCIdatashowthatthepresentedmethodcansearchquickertooptimizeprojectiondirectionthanGeneticAlgorithm(GA)andithasbetterclusteringresultscomparedwithtraditionallineardimensionreductionmethodforPrincipleComponentAnalysis(PCA).

简介：在里面三维(3D)轮廓测量，阶段移动profilometry(PSP)方法是广泛地使用的大多数一个。然而，因为多重设计，PSP的测量速度是很低的。以便改进测量速度，颜色栅栏条纹在这份报纸被用于测量。在测量期间，一仅仅颜色正弦曲线穗在测量目标上被投射。因此，测量速度极大地被改进。自从正在联合或在邻近的颜色栅栏条纹之间的干扰现象，一个颜色修正方法被用来改进测量结果。在修正在修正前比那好以后，为改正测量系统的非线性的错误的一个方法在这份报纸，和获得的图象的正弦曲线性质被建议。试验性的结果证明与这些修正方法，测量错误能被减少。因此，它能为高精确的3D重建支持一个好基础。

简介：多重信号分类(音乐)的表演关于仔细解决到达(DOA)的spaced方向的算法强烈取决于signal-to-noise比率(SNR)和快照。以便解决这个问题，由重建的一个方法有两噪音subspace的空间光谱函数和信号subspace在这份报纸被介绍。关键想法是使用在协变性矩阵包含的完整的信息并且改变分别地由他们的修订特征值在噪音和信号subspace上驾驶向量的设计重量。与MUSIC算法作比较，它也，并且显著地不增加任何计算复杂性，它有的优点同时减少噪音并且在低SNR和小样品下面保留高分辨率的能力大小的情形。模拟和实验结果被包括表明建议算法的优异表演。

简介：CognitiveRadio(CR)isapromisingtechniqueforthenextgenerationmobilecommuni-cationsystemforitscapabilitytosolvetheconflictsbetweenthescarcityandunderutilizationofspectrum.Inthispaper,aimingatmaximizingthesystemcapacityofamulti-antennaCRsystemonthepremisethatavoidinterferencetotheprimarysysteminthesamebandsimultaneously,aresourceallocationmethodwhichisabletoavoidinterferencebetweenPRimary(PR)andCRusersbypro-jectingthetransmitsignalsofCRusersonthenullspaceofthePRusers'channelsisproposed.CRuserswithbetterchannelconditionareselected,andtheinterferencefromCRsystemtoPRuserscanberemovedcompletelybyprojectingthetransmitsignalsofCRsystemonthenull-spaceofPRusers'channels.Parallelsub-channelsareconstructedforCRusersthroughSingularValueDecomposition(SVD).Atlast,waterfillingisalsoadoptedtoincreasetheCRusers'capacity.Simulationresultdemonstratesthatcomparedwithexistingmethods,ourmethodcanimprovetheachievablesumrateofCRusersaswellasreducetheoutageprobabilityofPRusers.

简介：MainstreamlineissignificantfortheYellowRiversituationforecastingandfloodcontrol.Aneffectivestatisticalfeatureextractionmethodisproposedinthispaper.Inthismethod,abetween-classscatteringmatrixbasedprojectionalgorithmisperformedtomaximizebetween-classdifferences,obtainingeffectivecomponentforclassification;thenhigh-orderstatisticsareutilizedasthefeaturestodescribethemainstreamlineintheprincipalcomponentobtained.Experimentsareperformedtoverifytheapplicabilityofthealgorithm.TheresultsbothonsynthesizedandrealscenesindicatethatthisapproachcouldextractthemainstreamlineoftheYellowRiverautomatically,andhasahighprecisioninmainstreamlinedetection.

简介：Ahybridmethodforsynthesizingantenna’sthreedimensional(3D)patternisproposedtoobtainthelowsidelobefeatureoftruncatedconeconformalphasedarrays.Inthismethod,theelementsoftruncatedconeconformalphasedarraysareprojectedtothetangentplaneinonegeneratrixofthetruncatedcone.Thentwodimensional(2D)Chebyshevamplitudedistributionoptimizationisrespectivelyusedintwomutualverticaldirectionsofthetangentplane.Accordingtothelocationoftheelements,theexcitationcurrentamplitudedistributionofeachelementontheconformalstructureisderivedreversely,thentheexcitationcurrentamplitudeisfurtheroptimizedbyusingthegeneticalgorithm(GA).Atruncatedconeproblemwith8×8elementsonit,anda3Dpatterndesiredsidelobelevel(SLL)upto35dB,isstudied.Byusingthehybridmethod,theoptimalgoalisaccomplishedwithacceptableCPUtime,whichindicatesthatthishybridmethodforthelowsidelobesynthesisisfeasible.