简介：树修剪是一个有效算法减少范围察觉(SD)的复杂性为多重输入多重产量(MIMO)通讯系统。怎么决定修剪统治的树，以及由在表演之间的折衷和复杂性能被完成的，仍然是一个开的问题。在这份报纸，修剪算法的一棵树基于最小的均方差(MMSE)被建议察觉。建议算法自从MMSE察觉的复杂性，第一预先形成MMSE察觉是很低的。然后，修剪的限制将被放根据MMSE答案的放大路径度量标准。复杂性和表演上的规模因素和他们的影响的选择也被讨论。通过分析和模拟，复杂性与可以忽略的表演降级和另外的计算显著地被减少，这被显示出。

简介：Intrusiondetectioncanbeessentiallyregardedasaclassificationproblem,namely,dis-tinguishingnormalprofilesfromintrusivebehaviors.Thispaperintroducesboostingclassificationalgorithmintotheareaofintrusiondetectiontolearnattacksignatures.Decisiontreealgorithmisusedassimplebaselearnerofboostingalgorithm.Furthermore,thispaperemploysthePrincipleCom-ponentAnalysis(PCA)approach,aneffectivedatareductionapproach,toextractthekeyattributesetfromtheoriginalhigh-dimensionalnetworktrafficdata.KDDCUP99datasetisusedintheseex-perimentstodemonstratethatboostingalgorithmcangreatlyimprovetheclassificationaccuracyofweaklearnersbycombininganumberofsimple“weaklearners”.Inourexperiments,theerrorrateoftrainingphaseofboostingalgorithmisreducedfrom30.2%to8%after10iterations.Besides,thispaperalsocomparesboostingalgorithmwithSupportVectorMachine(SVM)algorithmandshowsthattheclassificationaccuracyofboostingalgorithmislittlebetterthanSVMalgorithm’s.However,thegeneralizationabilityofSVMalgorithmisbetterthanboostingalgorithm.

简介：Themeasurementtechniquesoffemtosecondspectroscopyareeffectivemethodtoinvestigateultrafastdynamics,theyarewidelyusedinthefieldsofphysics,chemistryandbiology.Inthispaper,theprinciple,experimentsetupandtheapproachestodealwiththeexperimentdatawerepresented.Thendifferentmeasurementtechniquessuchastransientabsorptionspectroscopy,photonechoes,opticalKerreffectanddegeneratefour-wavemixingwereexplainedwithspecialexamples.Atlast,theapplicationprospectofmeasurementtechniquesoffemtosecondspectroscopywasforecasted.

简介：Modernnetworksystemshavemuchtroubleinsecurityvulnerabilitiessuchasbufferoverflow,bugsinMicrosoftInternet,sensornetworkroutingprotocoltoosimple,securityflawsofapplications,andoperatingsystems.Moreover,wirelessdevicessuchassmartphones,personaldigitalassistants(PDAs),andsensorshavebecomeeconomicallyfeasiblebecauseoftechnologicaladvancesinwirelesscommunicationandmanufacturingofsmallandlow-costsensors.Therearetypologiesofvulnerabilitiestobeexploitedinthesedevices.Inordertoimprovesecurities,manymechanismsareadopted,includingauthentication,cryptography,accesscontrol,andintrusiondetectionsystems(IDS).Ingeneral,intrusiondetectiontechniquescanbecategorizedintotwogroups:misusedetectionandanomalydetection.Themisusedetectionsystemsusepatternsofwell-knownattacksorweakspotsofthesystemstoidentifyintrusions.Theweaknessofmisusedetectionsystemsisunabletodetectanyfuture(unknown)intrusionuntilcorrespondingattacksignaturesareintrudedintothesignaturedatabase.Anomalydetectionmethodstrytodeterminewhetherthedeviationisfromtheestablishednormalusagepatternsornot.Thecriticalsuccessofanomalydetectionreliesonthemodelofnormalbehaviors.

简介：Oneofthemainrequirementsofcognitiveradiosystemsistheabilitytodetectthepresenceoftheprimaryuserwithfastspeedandpreciseaccuracy.Toachievethat,apossibletwo-stagespectrumsensingschemeissuggestedinthispaper.Morespecifically,afastspectrumsensingalgorithmbasedontheenergydetectionisintroducedfocusingonthecoarsedetection.Acomplementaryfinespectrumsensingalgorithmadoptsone-ordercyclostationarypropertiesofprimaryuser'ssignalsintimedomain.Sincetheone-orderfeaturedetectionisperformedintimedomain,thereal-timeoperationandlow-computationalcomplexitycanbeachieved.Also,itdrasticallyreduceshardwareburdensandpowerconsumptionasopposedtotwo-orderfeaturedetection.Thesensingperformanceoftheproposedmethodisstudiedandtheanalyticalperformanceresultsaregiven.Theresultsindicatethatbetterperformancecanbeachievedinproposedtwo-stagesensingdetectioncomparedtotheconventionalenergydetector.

简介：Theautomaticdetectionoffacesisaveryimportantproblem.Theeffectivenessofbiometricauthenticationbasedonfacemainlydependsonthemethodusedtolocatethefaceintheimage.Thispaperpresentsahybridsystemforfacesdetectioninunconstrainedcasesinwhichtheillumination,pose,occlusion,andsizeofthefaceareuncontrolled.Todothis,thenewmethodofdetectionproposedinthispaperisbasedprimarilyonatechniqueofautomaticlearningbyusingthedecisionofthreeneuralnetworks,atechniqueofenergycompactionbyusingthediscretecosinetransform,andatechniqueofsegmentationbythecolorofhumanskin.Awholeofpictures(facesandnofaces)aretransformedtovectorsofdatawhichwillbeusedforlearningtheneuralnetworkstoseparatebetweenthetwoclasses.Discretecosinetransformisusedtoreducethedimensionofthevectors,toeliminatetheredundanciesofinformation,andtostoreonlytheusefulinformationinaminimumnumberofcoefficientswhilethesegmentationisusedtoreducethespaceofresearchintheimage.Theexperimentalresultshaveshownthatthishybridizationofmethodswillgiveaverysignificantimprovementoftherateoftherecognition,qualityofdetection,andthetimeofexecution.

简介：Thispaperpresentsahumandetectionsysteminavision-basedhospitalsurveillanceenvironment.Thesystemiscomposedofthreesubsystems,i.e.backgroundsegmentationsubsystem(BSS),humanfeatureextractionsubsystem(HFES),andhumanrecognitionsubsystem(HRS).ThecodebookbackgroundmodelisappliedintheBSS,thehistogramoforientedgradients(HOG)featuresareusedintheHFES,andthesupportvectormachine(SVM)classificationisemployedintheHRS.Bymeansoftheintegrationofthesesubsystems,thehumandetectioninavision-basedhospitalsurveillanceenvironmentisperformed.Experimentalresultsshowthattheproposedsystemcaneffectivelydetectmostofthepeopleinhospitalsurveillancevideosequences.

简介：Whenthereturnscomefromtwoormoreunresolvedtargets(thesignalsarenotresolvedinthefrequencyortimedomains)inamonopulseradarsystem,thedirection-of-arrival(DOA)estimateindicatedbythemonopulseratioisnotthetrueinformationofthetrackedtarget.Thenthetrackingsystemsisinfluenced.Anapproachhasbeenproposedtodetectwhetherthereturnscomefromasingletargetortwounresolvedtargets.Theaboveapproachisextendedfromtwotothreeunresolvedtargets.Thesimulationindicatesthatthedetectionprobabilityunderthethreeunresolvedtargetsisnotsuretoexceedthedetectionprobabilityunderthetwounresolvedtargets.

简介：Forfacedetectionundercomplexbackgroundandillumination,adetectionmethodthatcombinestheskincolorsegmentationandcost-sensitiveAdaboostalgorithmisproposedinthispaper.First,byusingthecharacteristicofhumanskincolorclusteringinthecolorspace,theskincolorareainYCbCrcolorspaceisextractedandalargenumberofirrelevantbackgroundsareexcluded;thenforremedyingthedeficienciesofAdaboostalgorithm,thecost-sensitivefunctionisintroducedintotheAdaboostalgorithm;finallytheskincolorsegmentationandcost-sensitiveAdaboostalgorithmarecombinedforthefacedetection.Experimentalresultsshowthattheproposeddetectionmethodhasahigherdetectionrateanddetectionspeed,whichcanmoreadapttotheactualfieldenvironment.

简介：Inordertoovercometheexistingdisadvantagesofofflinelasershockpeeningdetectionmethods,anonlinedetectionmethodbasedonacousticwavesignalsenergyisprovided.Duringthelasershockpeening,anacousticemissionsen-soratadefinedpositionisusedtocollecttheacousticwavesignalsthatpropagateintheair.Theacousticwavesignalissampled,stored,digitallyfilteredandanalyzedbytheonlinelasershockpeeningdetectionsystem.Thenthesystemgetstheacousticwavesignalenergytomeasurethequalityofthelasershockpeeningbyestablishingthecorrespondencebetweentheacousticwavesignalenergyandthelaserpulseenergy.ThesurfaceresidualstressesofthesamplesaremeasuredbyX-raystressanalysisinstrumenttoverifythereliability.Theresultsshowthatboththesurfaceresidualstressandacousticwavesignalenergyareincreasedwiththelaserpulseenergy,andtheirgrowthtrendsareconsistent.Finally,theempiricalformulabetweenthesurfaceresidualstressandtheacousticwavesignalenergyisestablishedbythecubicequationfitting,whichwillprovideatheoreticalbasisforthereal-timeonlinedetectionoflasershockpeening.

简介：Thispaperpresentsafault-detectionmethodbasedonthephasespacereconstructionanddataminingapproachesforthecomplexelectronicsystem.TheapproachforthephasespacereconstructionofchaotictimeseriesisacombinationalgorithmofmultipleautocorrelationandΓ-test,bywhichthequasi-optimalembeddingdimensionandtimedelaycanbeobtained.Thedataminingalgorithm,whichcalculatestheradiusofgyrationofunit-masspointaroundthecentreofmassinthephasespace,candistinguishthefaultparameterfromthechaotictimeseriesoutputbythetestedsystem.Theexperimentalresultsdepictthatthisfaultdetectionmethodcancorrectlydetectthefaultphenomenaofelectronicsystem.

简介：动态察觉基于光学传感器和变化雷达是一个新方法检测飞行帮助灯的发光紧张。光学传感器能得到每盏灯的照明信息，变化雷达得到距离信息，然后数据合并技术被用来计算每盏灯的发光紧张。修改这个动态察觉系统的错误的一个方法被介绍。它由基于光学传感器使用山峰价值察觉估计每盏灯的精确位置避免累积错误和测量搬运人的旅行错误，然后修改光的纵向距离信息和横向的位置信息。察觉和变化系统的表演被一些实验验证并且出现在图画。

简介：Compressivesensingisarevolutionaryideaproposedrecentlytoachievemuchlowersamplingrateforsignals.Intheimageapplicationwithlimitedresourcesthecameradatacanbestoredandprocessedincompressedform.Analgorithmformovingobjectandregiondetectioninvideousingacompressivesamplingisdeveloped.Thealgorithmestimatesmotioninformationofthemovingobjectandregionsinthevideofromthecompressivemeasurementsofthecurrentimageandbackgroundscene.Thealgorithmdoesnotperforminversecompressiveoperationtoobtaintheactualpixelsofthecurrentimagenortheestimatedbackground.Thisleadstoacomputationallyefficientmethodandasystemcomparedwiththeexistingmotionestimationmethods.Theexperimentalresultsshowthatthesamplingratecanreduceto25%withoutsacrificingperformance.

简介：Asimpleinstrumentforthereal-timemeasurementofalgaeconcentrationandmappingdescribed.Theinstrumentusesapulsedshortarcxenonflashlampastheexcitedlightsources.Boththeexcitinglightandthefluorescencefromalgaechlorophyllaretransmittedalongafiberbundle.Themeasurementsensivitivityisanalyzedandtheexperimentresultisgiven.Theinstrumentispracticaltoin-situmeasurementatsea.

简介：Visualfiredetectiontechnologiescandetectfireandalarmwarningsearlierthanconventionalfiredetectors.Thisstudyproposesaneffectivevisualfiredetectionmethodthatcombinesthestatisticalfirecolormodelandsequentialpatternminingtechnologytodetectfireinanimage.Furthermore,theproposedmethodalsosupportsreal-timefiredetectionbyintegratingadaptivebackgroundsubtractiontechnologies.Experimentalresultsshowthattheproposedmethodcaneffectivelydetectfireintestimagesandvideos.ThedetectionaccuracyoftheproposedhybridmethodisbetterthanthatofCelik’smethod.

简介：交通灯察觉和识别为在城市的环境的自治开车是必要的。一个照相机基于算法因为即时柔韧的交通灯察觉和识别被建议，并且特别为自治车辆设计了。尽管当前的可靠交通灯识别算法操作在进行中的井，他们中的大多数主要在一个固定位置为察觉被设计，在真实世界的条件下面的自治车辆上的效果仍然是有限的。一些方法在自治车辆上完成高精确性，但是没有高精确的priori地图的帮助，他们不能通常工作。作者为这个问题介绍了一个基于照相机的算法。处理流动的图象能被划分成三步，包括预处理，察觉和识别。第一，red-green-blue(红绿蓝)颜色空间作为预处理的主要内容被变换成hue-saturation-value(HSV)。在察觉步，同时，先验的颜色阀值方法被用于起始的过滤优先的知识被执行扫描景色以便快速建立候选人区域。为识别，面向的坡度(公猪)的这张文章使用直方图展示并且也支持向量机器(SVM)认出交通灯的状态。我们的自治车辆上的建议系统被评估。与投票的计划，建议罐头在城市的enviroment为自治车辆提供足够的精确性。

简介：Inmausoleummurals,existingbubblesareonekindofthemostharmfuldefectsfortherepairandprotectionofrelics.Forthisreason,itisnecessarytodetectbubbles,especiallytheoneswithsmallsize.Amethodtodetectthesmallbubbleswithenhancedterahertz(THz)imagesisproposed.Tosimulatethebubblesinthemausoleummurals,circulargrooveshavebeenhiddenintheplasterandthenmeasuredbytheTHzreflectedtimedomainspectroscopyimagingsystem.Toobservethesmallbubblesinmurals,acomprehensiveenhancementalgorithmisadoptedtoprocesstheobtainedTHzimages.Withtheenhancedmethod,thecirculargroovesinthemuralscanbeobservedclearly,evenforthecirculargroovewithadiameterof1.5mm.Theresultsindicatethattheproposedcomprehensivemethodcanbeusedtodetectthetinydefectsofmurals.

简介：Whentheclassicalconstantfalse-alarmrate(CFAR)combinedwithfuzzyC-means(FCM)algorithmisappliedtotargetdetectioninsyntheticapertureradar(SAR)imageswithcomplexbackground,CFARrequiresblock-by-blockestimationofcluttermodelsandFCMclusteringconvergestolocaloptimum.Toaddresstheseproblems,thispaperpro-posesanewdetectionalgorithm:knowledge-basedcombinedwithimprovedgeneticalgorithm-fuzzyC-means(GA-FCM)algorithm.Firstly,thealgorithmtakestargetregion’smaximumandaverageintensity,area,lengthoflongaxisandlong-to-shortaxisratiooftheexternalellipseasfactorswhichinfluencethetargetappearingprobability.Theknowledge-baseddetectionalgorithmcanproducepreprocessresultswithouttheneedofestimationofcluttermodelsasCFARdoes.AfterwardtheGA-FCMalgorithmisimprovedtoclusterpre-processresults.IthasadvantagesofincorporatingglobaloptimizingabilityofGAandlocaloptimizingabilityofFCM,whichwillfurthereliminatefalsealarmsandgetbetterresults.TheeffectivenessoftheproposedtechniqueisexperimentallyvalidatedwithrealSARimages.

简介：DespiteextensiveresearchonR-trees,mostoftheproposedschemeshavenotbeenintegratedintoexistingDBMSowingtothelackofprotocolsofconcurrencycontrol.R-linktreeisanacceptabledatastructuretodealwiththisissue,butproblemslikephantomstillexist.Inthispaper,wefocusonaconflictdetectionschemebasedonR-linktreeforcompleteconcurrencycontrol.Anin-memoryoperationcontrollistisdesignedtosuspendconflictingoperations.Themainfeaturesofthisapproachare(1)itcanbeimplementedeasilyanddoesnotneedanyextrainformation;(2)Nodeadlocksareinvolvedinlockingscheme;(3)Non-conflictingoperationsarenotrestricted;and(4)PhantomproblemsinR-linktreeareavoidedthroughbeforehandpredication.Theexperimentresultsshowthatthisschemeiscorrectandgainsbettersystemperformance.

简介：Anoveladaptiveswitchingfilter(ASF)basedondirectionaldetectionisproposedfordenoisingtheimagesthatarehighlycorruptedbyimpulsenoise.Theproposedalgorithmemploysanefficientnoisedetectionmechanism.Itfirstemploysanefficientmethodtoestimatethedifferencesbetweenthecurrentpixelanditsneighborsalignedwith28directions.Thecurrentnoisepixelisreplacedbyamedianorameanvaluewithinanadaptivefilterwindowwithrespecttodifferentnoisedensities.Experimentalresultsshowthattheproposedapproachcannotonlyachieveverylowmiss-detectionratioandfalse-alarmratioevenuptohighnoisecorruption,butalsopreservethedetailedinformationofanimageverywell.