简介：Equivalentsourcelayer(ESL)imagingisanimportantkindofhigh-resolutionelectroencephalogram(EEG)imaging.Itconsistsoftwocategories:equivalentdipolelayer(EDL)andequivalentchargelayer(ECL).Bothofthemareassumedtobelocatedonornearthecorticalsurfaceandhavebeenproposedashigh-resolutionimagingmodalitiesorasintermediatestepstoestimatetheepicorticalpotential.Here,EDLandECLbasedonarealisticheadmodelarepresented,bothsimulationsandrealdataexperimentaredonetocomparethesetwomodels.TheresultsshowthatECLcanprovidehigherspatialresolutionaboutsourcelocationthanEDLdoes.

简介：纸为同时的图象分类和注解建议一个新奇概率的生产模型。模型考虑范畴信息能为图象注解提供珍贵信息的事实。一旦一幅图象的范畴被查明，注解词的范围能被缩小，并且产生无关的注解词的概率能被减少。到这个目的，根据班注解图象的想法在模型被介绍。用变化方法，模型的近似推理和参数评价算法被导出，并且为分类并且注解新图象的有效近似也被给。我们的模型的力量在二真实世界数据集上被表明：1600图象LabelMe数据集和1791图象UIUC运动数据集。实验结果证明分类表演与几个最先进的分类模型同样，当注解表演比几个最先进的注解模型的好时。

简介：音调模型(TM)集成是为官员语音识别的一项重要任务。有效有区别地使用当把TM分数集成到多传递语音识别时，放大因素训练被证明了。而且，可伸缩的上下文依赖者(CD)能更好被申请在模型之间的插值。这条途径的一限制是很多参数将被介绍，它使这种技术容易到训练过度。在这份报纸，我们建议由使用自动地导出的语音的决定树导致上下文依赖者模型重量。在每个树节点的问题被选择在训练数据上最小化期望的识别错误。首先订客观功能被用于问题修剪使树造有效的最小的电话错误(MPE)的近似。连续官员语音识别上的试验性的结果证明方法能够导致最关键的语音的上下文并且与少得多的参数获得重要错误减小，与由使用放大参数的手工地设计的上下文依赖者获得了那相比。

简介：Anadaptivefuzzyslidingmodecontrol(AFSMC)approachisproposedforaroboticairship.First,themathematicalmodelofanairshipisderivedintheformofanonlinearcontrolsystem.Second,anAFSMCapproachisproposedtodesigntheattitudecontrolsystemofairship,andtheglobalstabilityoftheclosed-loopsystemisprovedbyusingtheLyapunovstabilitytheorem.Finally,simulationresultsverifytheeffectivenessandrobustnessoftheproposedcontrolapproachinthepresenceofmodeluncertaintiesandexternaldisturbances.

简介：CooperativepathdynamicplanningofaUCAV(unmannedcombatairvehicle)teamnotonlyconsidersthecapabilityoftaskrequirementofsingleUCAV,butalsoconsidersthecooperativedynamicconnectionamongmembersoftheUCAVteam.AcooperativepathdynamicplanningmodeloftheUCAVteambyapplyingaglobaloptimizationmethodisdiscussedinthispaperandthecorrespondingmodelisbuiltandanalyzed.Bytheexamplesimulation,thereasonableresultacquiredindicatesthatthemodelcouldmeetdynamicplanningdemandunderthecircumstanceofmembershipfunctions.Themodeliseasytoberealizedandhasgoodpracticability.

简介：Anovelactivecontourmodelisproposed,whichincorporateslocalinformationdistributionsinafuzzyenergyfunctiontoeffectivelydealwiththeintensityinhomogeneity.Moreover,theproposedmodelisconvexwithrespecttothevariablewhichisusedforextractingthecontour.Thismakesthemodelindependentontheinitialconditionandsuitableforanautomaticsegmentation.Furthermore,theenergyfunctionisminimizedinacomputationallyefficientwaybycalculatingthefuzzyenergyalterationsdirectly.Experimentsarecarriedouttoprovetheperformanceoftheproposedmodeloversomeexistingmethods.Theobtainedresultsconfirmtheefficiencyofthemethod.