基于OpenCV的人脸检测-毕设论文

基于OpenCV的人脸检测摘要人脸检测是指使用计算机在动态的场景或复杂的背景中检测出人脸的存在，并确定所有人脸的位置与大小的过程。人脸检测技术不仅是人脸识别、表情识别、人脸跟踪等技术的重要前提条件，同时在模式识别、人机交互、智能监控、视频检索等领域也引起了广泛的重视。本论文简单介绍了国内外人脸识别技术研究及应用的发展现状及其难点分析。在第二章中介绍了包括基于知识的方法、特征不变量方法、模板匹配方法、基于外观的方法四类检测方法；在第三章简单介绍了四种经典的检测方法，包括特征脸、神经网络、隐马尔可夫模型方法、支持向量机。在第四章重点分析了AdaBoost算法中集成机器学习的一个重要机制：多个弱分类器集成的方法，机器学习中的弱学习到强学习，集成的关键是投票，最简单的方法是“绝对多数”的方法。在第五章详细分析了AdaBoost算法检测速度快、可以检测任意尺度的图像的特点。在这个理论基础上，本文中人脸算法的研究基于OpenCV开源代码，在OpenCV开源代码中设计了一些基础的数据类型和一些帮助数据类型。由于OpenCV的源代码完全开放，本文的研究中利用这套代码在PC上以VisualC++集成开发环境做平台搭建了一个基于OpenCV的人脸检测系统，进行了人脸检测仿真，并对其结果进行分析。关键词：人脸检测；AdaBoost；分类器；OpenCV2FaceDetectionBasedonOpenCVAbstractHumanfacedetectionmeansthatforagivenimageorvideo，todeterminewhetheritcontainsfaceregions，ifso，determinesthenumber,theexactlocationandthesizeofallthefaces．Humanfacedetectionisnotonlyanecessarypreconditionoffacerecognition，expressionrecognitiontechnology,facetracking，butalso，itplaysailimportantroleinapplicationslikeintheintelligenthuman-computerinteraction，videoconferencing，intelligentsurveillance，videoretrievalandsoon．Therefore，facedetectiontechnologyattractedwidespreadattentioninpatternrecognition，computervision，human-computerinteractionandotherfields．Thispaperdescribesthenationalandinternationalrecognitiontechnologyapplicationsresearchanddevelopmentandanalysisthedifficultsituation.InthesecondchaptertoincludeKnowledge-basedMethods,FeatureInvariantApproaches,TemplateMatchingMethods,Appearance-basedMethods.Inthethirdchapterdescribesthefourclassicaldetectionmethods,includingEigenface,ArtificialNeuralNetwork,HiddenMarkovModel,SupportVectorMachines.Inchapter4ontheanalysisoftheAdaBoostalgorithmsintegrationofmachinerytostudyofanimportantmechanism:moreintegratedapproachintheclassificationofthestudyofstudyatthelearningandintegrationisthekeytothevote,thesimplestmethodistheabsolutemajority.InchapterfivedetailedanalysisAdaBoostalgorithmtestspeed,youcandetectanyscaletheimagequality.Thispaper’sresearchisbasedontheOpenCVsourcecode，inwhichsomebasicdatatypeandhelpingdatatypewerecreated，andbecauseoftheopermessofthecode，webuildahumanfacedetectionsystemintheVisualC++environment.Keywords:facedetection;AdaBoost;classifier;openCV3目录摘要...........................................................................................................................................................1Abstract..........................................................................................................................................................2第1章人脸检测........................................................................................................................................51.1背景..............................................................................................................................................51.2目前的研究状况..........................................................................................................................61.3概念..............................................................................................................................................61.4人脸检测及其合成技术的应用领域..........................................................................................61.5人脸检测评价标准......................................................................................................................7第2章检测方法........................................................................................................................................92.1基于知识的方法..........................................................................................................................92.2特征不变量方法..........................................................................................................................92.3模板匹配方法............................................................................................................................102.4基于外观的方法........................................................................................................................11第3章经典方法概述..............................................................................................................................133.1特征脸........................................................................................................................................133.2神经网络....................................................................................................................................133.3隐马尔可夫模型方法................................................................................................................143.4支持向量机................................................................................................................................14第4章Adaboost算法.............................................................................................................................164.1概述............................................................................................................................................164.1.1Adaboost算法简介........................................................