搜索
摘要模式识别就是机器识别、计算机识别或者机器自动化识别，目的在于让机器自动识别事物，使机器能做以前只能由人类才能做的事，具备人所具有的对各种事物与现象进行分析、描述与判断的部分能力。它研究的目的就是利用计算机对物理对象进行分类，在错误概率最小的条件下，使识别的结果尽量与客观事物相符合。随着人们对人工神经网络的不断地认识，神经网络是指用大量的简单计算单元构成的非线性系统,它在一定程度和层次上模仿了人脑神经系统的信息处理、存储及检索功能，因而具有学习、记忆和计算等智能处理功能。这样人们利用人工神经网络具有高度的并行性，高度的非线性全局作用以及良好的容错性与联想记忆功能，并且具有良好的自适应、自学习功能等突出特点，可运用MATLAB神经网络工具箱中的神经网络模型，对经过训练的神经网络可以有效地提取信号、语音、图像等感知模式的特征，并能解决现有启发式模式识别系统不能很好解决的不变量探测、抽象和概括等问题。这样神经网络可应用于模式识别的特征提取、聚类分析、边缘检测、信号增强以及噪声抑制、数据压缩等各个环节。使用机器来进行模式的识别是一项非常有用的工作，能够辨别符号等系列的机器是很有价值的。目前，模式识别技术可以应用指纹识别、IC卡技术应用、字符识别等实例。模式识别成为人工神经网络特别适宜求解的一类问题。因此，神经网络技术在模式识别中也得到广泛应用与发展。关键词：模式识别；人工神经网络；神经网络模型；神经网络技术AbstractPatternRecognitionisthemachineidentification,computeridentificationoridentificationofmachineautomation,machineaimedatautomaticidentificationofthingstodobeforethemachinecanonlybemadebymancando,withpeoplewithallkindsofthingsandonananalysisofthephenomenon,describedwiththeabilitytodeterminethepart.Itisthepurposeofthestudyofthephysicalobjecttousethecomputerforclassification,theprobabilityofthesmallestinthewrongconditions,sothattheresultsofrecognitionasfaraspossibleinlinewithobjectivethings.Asartificialneuralnetworktorecognizethecontinuing,neuralnetworkreferstoalargenumberofsimplecalculationunitconsistingofnon-linearsystem,whichtosomeextentandlevelsystemtoimitatethehumanbrain'sinformationprocessing,storageandretrievalfunctions,whichhaslearning,memoryandcomputingfunctionssuchasintelligentprocessing.Suchpeopletouseartificialneuralnetworkwithahighdegreeofparallelism,theoverallroleofahighdegreeofnon-linearandgoodfaulttoleranceandassociativememoryfunction,andhavegoodself-adaptive,self-learningfunction,suchasprominentfeatures,theavailabilityofMATLABneuralnetworktoolboxTheneuralnetworkmodeltrainedneuralnetworkcaneffectivelyextractthesignal,voice,videoandotherfeaturesofperceptualpatternsandheuristicstosolvetheexistingpatternrecognitionsystemsarenotwellresolvedinvariantdetection,suchasabstractandsummaryissues.Thisneuralnetworkpatternrecognitioncanbeappliedtofeatureextraction,clusteringanalysis,edgedetection,signalenhancementandnoisesuppression,datacompression,suchasvariouslinks.Theuseofmachinesforpatternrecognitionisaveryusefulwork,suchasseriesofsymbolstoidentifythemachinesareofgreatvalue.Atpresent,thepatternrecognitiontechnologycanbeappliedtofingerprintidentification,ICcardtechnologyapplications,suchasexamplesofcharacterrecognition.Artificialneuralnetworkpatternrecognitionhasbecomeespeciallysuitableforsolvingaclassofproblem.Therefore,theneuralnetworkpatternrecognitiontechnologyisalsowidelyusedanddevelopment.Keywords：patternrecognition；artificialneuralnetwork；neuralnetworkmodel；neuralnetworktechnology目录引言..........................................................................................................................................................-1-1模式识别概述.............................................................................................................................-1-1.1模式识别基本概念..........................................................................................................................-1-1.2模式识别系统....................................................................................................................................-2-1.3模式识别的主要方法.....................................................................................................................-2-1.4模式识别应用....................................................................................................................................-3-2人工神经网络概述..................................................................................................................-4-2.1人工神经元模型...............................................................................................................................-4-2.1.1阀值函数、Sigmoid函数和分线段性函数..................................................................-5-2.2人工神经网络模型..........................................................................................................................-5-2.3神经网络学习特点..........................................................................................................................-6-2.4人工神经网络在模式识别问题中应用优势..........................................................................-6-3神经网络模式识别..................................................................................................................-7-3.1基于BP神经网络的模式识别.....................................................................................................-7-3.1.1BP神经网络模型简述............................................................................................................-7-3.1.2BP学习算法...............................................................................................................................-8-3.1.3BP神经网络应用于字符识别..............................................................................................-9-3.2基于径向基函数神经网络的模式识别..................................................................................-16-3.2.1径向基函数神经网络模型简述...........................