简介：Thispaperconcernstheproblemofobjectsegmentationinreal-timeforpickingsystem.Aregionproposalmethodinspiredbyhumanglancebasedontheconvolutionalneuralnetworkisproposedtoselectpromisingregions,allowingmoreprocessingisreservedonlyfortheseregions.Thespeedofobjectsegmentationissignificantlyimprovedbytheregionproposalmethod.Bythecombinationoftheregionproposalmethodbasedontheconvolutionalneuralnetworkandsuperpixelmethod,thecategoryandlocationinformationcanbeusedtosegmentobjectsandimageredundancyissignificantlyreduced.Theprocessingtimeisreducedconsiderablybythistoachievetherealtime.Experimentsshowthattheproposedmethodcansegmenttheinterestedtargetobjectinrealtimeonanordinarylaptop.

简介：Synchronouschipsealisanadvancedroadconstructingtechnology,andthegravelcoveragerateisanimportantindicatoroftheconstructionquality.Inthispaper,anovelapproachforgravelcoverageratemeasurementisproposedbasedondeeplearning.Convolutionalneuralnetwork(CNN)isusedtosegmenttheimageofgroundcoveredwithgravels,andthegravelcoveragerateiscomputedbythepercentageofgravelpixelsinthesegmentedimage.Thegravelcoverageratedatasetformodeltrainingandtestingisbuilt.Theperformanceoffullyconvolutionalneuralnetwork(FCN)andU-Netmodelinthedatasetistested.AbettermodelnamedGravelNetisconstructedbasedonU-Net.Thescaledexponentiallinearunit(SELU)isemployedintheGravelNettoreplacethepopularcombinationofrectifiedlinearunit(ReLU)andbatchnormalization(BN).Dataaugmentationandalphadropoutareperformedtoreduceoverfitting.Theexperimentalresultsdemonstratetheeffectivenessandaccuracyofourproposedmethod.OurtrainedGravelNetachievesthemeangravelcoveragerateerrorof0.35%ontestdataset.

简介：ThemaingoalofroutingsolutionsistosatisfytherequirementsoftheQualityofService(QoS)foreveryadmittedconnectionaswellastoachieveaglobalefficiencyinresourceutilization.InthispaperproposesasolutionbasedonHopfieldneuralnetwork(HNN)todealwithoneofrepresentativeroutingproblemsinuni-castrouting,i.e.themulti-constrained(MC)routingproblem.ComputersimulationshowsthatwecanobtaintheoptimalpathveryrapidlywithournewLyapunovenergyfunctions.

简介：Basedoncurrentresearchonapplicationsofchaoticneuronnetworkforinformationprocessing,thestabilityandconvergenceofchaoticneuronnetworkareprovedfromtheviewpointofenergyfunction.Moreover,anewauto-associativematrixisdevisedforartificialneuralnetworkcomposedofchaoticneurons,thus,animprovedchaoticneuronnetworkforassociativememoryisbuiltup.Finally,theassociativerecallingprocessofthenetworkisanalyzedindetailandexplanationsofimprovementaregiven.

简介：AsimplenewBPalgorithmnamedcircleBPalgorithmisintroduced.Withthisalgorithm,localminimumscanbecompletelygotridofandlearningspeedcanimprovedramatically.ItcanbeeasilydesignedintothecircuitryandadvancefurthertheapplicationofMLPneuralnetwork.

简介：Gatematrixlayoutproblemplaysanimportantroleinintegratedcircuitdesign,butitsoptimizationisNP-hard.Inthispaper,typicalgatelayoutproblemisanalysedandadaptedtoneuralnetworkrepresentation,furthermorethesimulatedresultsaregiven.

简介：ByuseofHopfieldmodelandbasissolutionofhomogeneouslinearequationswhichareestablishedinaccordancewithconsistentstate,apracticaldecisionmethodfortheexistenceofoptimalHopfieldmodelofcombinationalcircuitsisprovided.Finally,anexampleisgiven.

简介：Inthisstudy,aMulti-LayerBPneuralnetwork(MLBP)withdynamicthresholdsisemployedtobuildaclassifiermodel.Astothedesignoftheneuralnetworkstructure,theoreticalguidanceandplentifulexperimentsarecombinedtooptimizethehiddenlayers'parameterswhichincludethenumberofhiddenlayersandtheirnodenumbers.Theclassifierwithdynamicthresholdsisusedtostandardizetheoutputforthefirsttime,anditimprovestherobustnessofthemodeltoahighlevel.Finally,theclassifierisappliedtoforecastboxofficerevenueofamoviebeforeitstheatricalrelease.ThecomparisonresultswiththeMLPmethodshowthattheMLBPclassifiermodelachievesmoresatisfactoryresults,anditismorereliableandeffectivetosolvetheproblem.

简介：

简介：Thispaperdescribesamodifiedspeed-sensorlesscontrolforinductionmotor(IM)basedonspacevectorpulsewidthmodulationandneuralnetwork.AnElmanANNmethodtoidentifytheIMspeedisproposed,withIMparametersemployedasassociatedelements.TheBPalgorithmisusedtoprovideanadaptiveestimationofthemotorspeed.Theeffectivenessoftheproposedmethodisverifiedbysimulationresults.TheimplementationonTMS320F240fixedDSPisprovided.

简介：Thispaperpresentsanewsolutiontotheimagesegmentationproblem,whichisbasedonfuzzy-neural-networkhybridsyste

简介：一个不规则的分割的区域编码算法基于脉搏，联合神经网络(PCNN)被介绍。PCNN有联合脉搏、可变的阀值，有近似灰色的价值的这些邻近的象素能同时通过被激活的性质。一个人能得出PCNN有的优点认识到地区性的分割，和细节的一个结论原来的图象能被分割图象的参数调整完成，并且同时，小分割的区域能被避免。为不规则的分割的区域的更好的近似，Gram-Schmidt方法，一组orthonormal基底函数被从一组线性独立的起始的底构造工作，被采用。因为重建的orthonormal方法，重构图象的质量能极大地被改进，进步图象传输将也是可能的。

简介：

简介：Choosingtherightcharacteristicparameteristhekeytofaultdiagnosisinanalogcircuit.Thefeatureevaluationandextractionmethodsbasedonneuralnetworkarepresented.Parameterevaluationofcircuitfeaturesisrealizedbytrainingresultsfromneuralnetwork;thesuperiornonlinearmappingcapabilityiscompetentforextractingfaultfeatureswhicharenormalizedandcompressedsubsequently.Thecomplexclassificationproblemonfaultpatternrecognitioninanalogcircuitistransferredintofeatureprocessingstagebyfeatureextractionbasedonneuralnetworkeffectively,whichimprovesthediagnosisefficiency.Afaultdiagnosisillustrationvalidatedthismethod.

简介：

简介：Handwrittensignaturerecognitionispresentedbasedonananglefeaturevectorbyusingtheartificialneuralnetwork(ANN)inthisresearch.Eachsignatureimagewillberepresentedbyananglevector.ThefeaturevectorwillconstitutetheinputtotheANN.Thecollectionofsignatureimagesisdividedintotwosets.OnesetwillbeusedfortrainingtheANNinasupervisedfashion.TheothersetwhichisneverseenbytheANNwillbeusedfortesting.Aftertraining,theANNwillbetestedbyrecognizingthesignatures.Whenasignatureisclassifiedcorrectly,itisconsideredcorrectrecognition,otherwiseitisafailure.Theachievedrecognitionrateofthissystemis94%.

简介：ＮｅｕｒａｌＮｅｔｗｏｒｋＳｉｇｎａｌＰｒｏｃｅｓｉｎｇＡｐｐｒｏａｃｈｆｏｒＤａｍａｇｅＡｓｓｅｓｓｍｅｎｔｉｎＦｉｂｅｒｏｐｔｉｃＳｍａｒｔＭａｔｅｒｉａｌＳｙｓｔｅｍｓａｎｄＳｔｒｕｃｔｕｒｅｓ①②ＴＵＹａｑｉｎｇ（Ｄｅｐａｒｔ．ｏｆＡｕｔｏｍ...

简介：

简介：Performancerobustnessproblemsviathestatefeedbackcontrollerareinvestigatedforaclassofuncertainnonlinearsystemswithtime-delayinbothstateandcontrol,inwhichtheneuralnetworksareusedtomodelthenonlinearities.Byusinganappropriateuncertaintydescriptionandthelineardifferenceinclusiontechnique,sufficientconditionsforexistenceofsuchcontrollerarederivedbasedonthelinearmatrixinequalities(LMIs).UsingsolutionsofLMIs,astatefeedbackcontrollawisproposedtostabilizetheperturbedsystemandguaranteeanupperboundofsystemperformance,whichisapplicabletoarbitrarytime-delays.

简介：Duetothedemandofdataprocessingforpolariceradarinourlaboratory,aCurveletThresholdingNeuralNetwork(TNN)noisereductionmethodisproposed,andanewthresholdfunctionwithinfinite-ordercontinuousderivativeisconstructed.ThemethodisbasedonTNNmodel.InthelearningprocessofTNN,thegradientdescentmethodisadoptedtosolvetheadaptiveoptimalthresholdsofdifferentscalesanddirectionsinCurveletdomain,andtoachieveanoptimalmeansquareerrorperformance.Inthispaper,thespecificimplementationstepsarepresented,andthesuperiorityofthismethodisverifiedbysimulation.Finally,theproposedmethodisusedtoprocesstheiceradardataobtainedduringthe28thChineseNationalAntarcticResearchExpeditionintheregionofZhongshanStation,Antarctica.Experimentalresultsshowthattheproposedmethodcanreducethenoiseeffectively,whilepreservingtheedgeoftheicelayers.