简介：Gesturerecognitionisanimportantresearchinthefieldofhuman-computerinteraction.HandGesturesarestrongvariableandflexible,sothegesturerecognitionhasalwaysbeenanimportantchallengefortheresearchers.Inthispaper,wefirstoutlinedthedevelopmentofgesturesrecognition,anddifferentclassificationofgesturesbasedondifferentpurposes.Thenwerespectivelyintroducedcommonmethodsusedintheprocessofgesturesegmentation,featureextractionandrecognition.Finally,thegesturerecognitionwassummarizedandthestudyingprospectsweregiven.

简介：

简介：蛋白质褶层结构是比它的氨基酸顺序保存并且仔细与生物功能联系了推测蛋白质结构的类似是在结构的生物学的一个基本问题并且在蛋白质褶层分类，褶层功能推理，和蛋白质结构起一个关键作用的更多预言。大进步为认为结构的类似被建议了在这块地和许多方法里在最近的年里被取得了，包括为蛋白质结构比较，从数据库的蛋白质结构的检索，和ligand绑定的方法地点比较。大多数那些方法能在万维网上是可得到的，但是所有方法的评估仍然是一个难问题。这篇论文为结构类似总结一些流行方法和最近的方法，包括结构排列，蛋白质结构检索，和ligand绑定地点排列。

简介：

简介：

简介：AnovelDiscreteWaveletTransform(DWT)basedHiddenMarkovModule(HMM)forfacerecognitionispresentedinthisletter.ToimprovetheaccuracyofHMMbasedfacerecognitionalgorithm,DWTisusedtoreplaceDiscreteCosineTransform(DCT)forobservationsequenceex-traction.Extensiveexperimentsareconductedontwopublicdatabasesandtheresultsshowthattheproposedmethodcanimprovetheaccuracysignificantly,especiallywhenthefacedatabaseislargeandonlyfewtrainingimagesareavailable.

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：Digitrecognitionfromanaturalscenetextinvideosurveillance/broadcastingapplicationsisachallengingresearchtaskduetoblurred,fontvariations,twisted,andnon-uniformcolordistributionissueswithadigitinanaturalscenetoberecognized.Inthispaper,tosolvethedigitnumberrecognitionproblem,aprincipal-axisbasedtopologycontourdescriptorwithsupportvectormachine(SVM)classificationisproposed.Thecontributionsofthispaperinclude:a)alocaldescriptorwithSVMclassificationfordigitrecognition,b)higheraccuracythanthestate-of-theartmethods,andc)lowcomputationalpower(0.03second/digitrecognition),whichmakethismethodadoptabletoreal-timeapplications.

简介：Theanalysisoftheradiatednoiseofvesselsgiveninthispapershowssomestrongsuperposedlinecomponentsinlowfrequencyspeetrumbelow100Hzoccurringatdiscretefrequencieswhichcorrespondwiththerotationspeedofpropellershaft,orpropellerbladefrequency,ortheirharmonicfre-quencies.sincethelinecomponentsreflectpropeller＇workingcharacteristics,thepropller＇sfeaturescanbeextracteddirectlyfromlow-frequencylinecom-ponentsinadditiontodemodulatedlinecomponent.Sotherearetwowaystoextractthefeatures,oneisdirectway,theotherisdemodulationway.Detec-tionperformanceofthelinecomponentinbackground-noiseisdiscussedinthispaper.ThesignallevelisdefinedasthedifrerencebetweenthePDF＇s(ProbabilityDensityFunction)meanofthepeakofthelinecomponentandPDF＇smeanorthebackground-noise.Indircetwaythesignallevelofthelinecomponentisproportionaltothesignalnoiseratio(S/N).Indemodulationwaythesignallevelofdemodulatedlinecomponent

简介：Thispaperdescribesanewkindofneuralnetwork-QuantumNeuralNetwork(QNN)anditsapplicationtorecognitionofcontinuousdigits.QNNcombinestheadvantagesofneuralmodelingandfuzzytheoreticprinciples.Experimentresultsshowthatmorethan15percenterrorreductionisachievedonaspeaker-independentcontinuousdigitsrecognitiontaskcomparedwithBPnetworks.

简介：

简介：Ahandgesturerecognitionmethodispresentedforhuman-computerinteraction,whichisbasedonfingertiplocalization.First,handgestureissegmentedfromthebackgroundbasedonskincolorcharacteristics.Second,featurevectorsareselectedwithequalintervalsontheboundaryofthegesture,andthengestures'lengthnormalizationisaccomplished.Third,thefingertippositionsaredeterminedbythefeaturevectors'parameters,andanglesoffeaturevectorsarenormalized.Finallythegesturesareclassifiedbysupportvectormachine.Theexperimentalresultsdemonstratethattheproposedmethodcanrecognize9gestureswithanaccuracyof94.1%.

简介：Inthispaper,weproposeafacerecognitionapproach-StructedSparseRepresentation-basedclassificationwhenthemeasurementofthetestsampleislessthanthenumbertrainingsamplesofeachsubject.Whenthisconditionisnotsatisfied,weexploitNearestSubspaceapproachtoclassifythetestsample.Inordertoadaptallthecases,wecombinethetwoapproachestoanadaptiveclassificationmethod-Adaptiveapproach.TheadaptiveapproachyieldsgreaterrecognitionaccuracythantheSRCapproachandCRC_RLSapproachwithlowsamplerateontheExtendYaleBdataset.Anditismoreefficientthanothertwoapproaches.

简介：Theaccuracyoffacealignmentaffectsgreatlytheperformanceofafacerecognitionsystem.Sincethefacealignmentisusuallyconductedusingeyepositions,thealgorithmforaccurateeyelo-calizationisessentialfortheaccuratefacerecognition.Inthispaper,analgorithmisproposedforeyelocalization.First,theproperAdaBoostdetectionisadaptivelytrainedtosegmenttheregionbasedonthespecialgraydistributionintheregion.Afterthat,afastradialsymmetryoperatorisusedtopre-ciselylocatethecenterofeyes.Experimentalresultsshowthatthemethodcanaccuratelylocatetheeyes,anditisrobusttothevariationsoffaceposes,illuminations,expressions,andaccessories.

简介：以便在实际侦察环境解决emitter识别问题，属性数学被介绍。属性集合和属性措施的基本概念和理论详细被描述。基于属性，措施在这篇论文被介绍的一个新属性识别方法。申请例子被给，它表明这个新方法精确、有效。而且，为认出emitter目的计算机摸拟;计算机模拟被选择，并且与通过模拟的古典统计模式识别相比。优秀试验性的结果证明这作为与存在统计模式识别技术相比是一个崭新的属性识别方法。

简介：Analgorithmforfacedescriptionandrecognitionbasedonmulti-resolutionwithmulti-scalelocalbinarypattern(multi-LBP)featuresisproposed.Thefacialimagepyramidisconstructedandeachfacialimageisdividedintovariousregionsfromwhichpartialandholisticlocalbinarypatter(LBP)histogramsareextracted.AllLBPfeaturesofeachimageareconcatenatedtoasingleLBPeigenvectorwithdifferentresolutions.ThedimensionalityofLBPfeaturesisthenreducedbyalocalmarginalignment(LMA)algorithmbasedonmanifold,whichcanpreservethebetween-classvariance.Supportvectormachine(SVM)isappliedtoclassifyfacialimages.ExtensiveexperimentsonORLandCMUfacedatabasesclearlyshowthesuperiorityoftheproposedschemeoversomeexistedalgorithms,especiallyontherobustnessofthemethodagainstdifferentfacialexpressionsandposturesofthesubjects.

简介：Withtherapiddevelopmentofbraincomputerinterface(simplycalledBCI),electroencephalography(EEG)willbeanotherinterestingbio-electricalsignalappliedinroboticsafterEMG.Inordertorealizeitfinally,theaccuratemeasurementandpatternrecognitionofEEGsignalmustbeaveryimportantandelementaryresearchobjective.Basedonourcurrentresearchesandsomereportsfromtheotherinternationalcolleaguesinthefield,wedeeplydiscussthebasiccharacteristicsofEEGsignal,thedevelopmentandselectionofEEGmeasurementsystem,featureextractionandrecognitionmethodsofEEGsignal,andthenreviewEEG'sapplicationsinroboticsaswellasthefutureresearchtrendsinthispaper.

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：Thearticleintroducesradontransformofimageaswellasamethodtocalculategeometricmomentsunderradontransform.Bymakinguseoftheanti-interferencecharacterofRadontransform,thispaperpresentsamethodofextractingimage'smomentfeature,whichcangetamomentfeaturematrixofimageunderRadontransform,aswellasamethodtorecognizeimagebyusingtheSVofthismatrix.

简介：Anewdigitalmodulationrecognitionalgorithmbasedontheinstantaneousinformationisproposedtoimprovetherecognitionsuccessrateinthelowsignalnoiseratio(SNR).Firstdenoisingoftheinstantaneousinformationisoptimizedbywaveletfilter,whichcanimprovetherecognitionabilityatlowSNR.Besidestheexisting3keyfeatureparameters,3newkeyfeatureparametersareproposedtobeusedasthedecisioncriteriaforidentifyingdifferenttypesofdigitalmodulation,whichsimplifiestherecognitionprocessandimprovestherecognitionabilityatlowSNR.Thesimulationsdemonstratethatallmodulationtypesofinteresthavebeenclassifiedwithsuccessrateofnolowerthan99%whenSNRis10dB.EveniftheSNRislowerthan5dB,thesuccessrateisover95.4%formostofthemodulationtypes.