人脸识别系统的研究与实现

I毕业设计（论文）人脸识别系统的研究与实现学院（系）：专业班级：学生姓名：指导教师：II摘要人脸识别技术是包括人脸检测技术和人脸身份认证技术在内的识别技术。人脸检测技术是根据所获得的视频或者图片信息，利用图像处理和计算机视觉图像算法，从图像中判断是否有人脸，并给出存在人脸的数量和位置；人脸身份认证技术是通过脸与脸的匹配识别人脸的身份。论文研究了人脸识别的基本原理和算法，分析了人脸识别系统的需求，设计了人脸识别系统的总体框架和主要功能模块，包括图像获取功能、图像预处理功能、人脸跟踪定位功能和人脸识别等。直接使用VS2010调用OpenCV库的视频捕获和图像处理函数获取图像并进行图像预处理，利用基于Adaboost算法的人脸检测算法进行人脸检测，并将人脸特征标记出来，实现了一个人脸检测系统。最后根据EHMM人脸识别算法，设计并实现了一个人脸识别软件系统，并结合已有的人脸数据库，对系统进行了实验仿真，实验结果表明本文设计的人脸识别系统是稳定的，并具有良好的人脸识别率。关键词：人脸识别；特征提取；C++；VS2013；图像预处理IIAbstractFacerecognitiontechnologyisatechnologywhichincludesfacedetectiontechnologyandfaceauthenticationtechnology.Facedetectiontechnologyisbasedonvideoorimageinformationobtainedtojudgewhetherthereisafaceintheimagebyusingimageprocessingandcomputervisionalgorithms,andshowthenumberandpositionoftheface.Faceauthenticationtechnologyisbyfacematchtorecognizetheidentityoftheface.Thebasicprinciplesandalgorithmsoffacerecognitionarestudiedinthedissertation,andthedemandforfacerecognitionsystemisanalyzed.Theoverallframeworkandthemainfunctionalmodulesoffacerecognitionsystemisdesigned,whichincludesimageacquisitionfunction,imagepreprocessingfunction,facetrackingandpositioningfunction,facerecongnitionfunctiongandsoon.Afacedetectionsystemisdeveloped,toimplementthereal-timevideofacedetectiontracking.ThebasicprinciplesoftheAdaboostfacedetectionalgorithmandtheHiddenMarkovModel-basedrecognitionalgorithmisdescribed.VS2010isusedtocallvideocaptureandimageprocessingfunctionofOpenCVtogetandpreprocesstheimage.AndthenfacedetectionalgorithmsbasedonAdaboostalgorithmisusedtodetectthehumanfaceandimplementedafacedetectionsystemtotrackandpositiontheface.Finally,designedandimplementatedafacerecognitionsoftwaresystemwiththeHMMfacerecognitionalgorithm.Combiningwiththeexistingexperimentalfacedatabase,simulatedthesystem.Experimentresultsshowedthatthedesignofthefacerecognitionsystemisstable,anditalsohasgoodrecognitionrates.Keywords:FaceRecognition;FeatureExtraction;HMMAlgorithm;ImagePre-processing;AdaboostAlgorithm目录第1章绪论...........................................................................................................................11.1人脸识别的研究背景和意义.......................................................................................................11.1.1研究背景...............................................................................................................11.1.2研究意义...............................................................................................................21.2国内外人脸识别技术研究现状..................................................................................................21.3本文所做的主要工作及论文内容安排....................................................................................31.3.1本文所做的主要工作...........................................................................................31.3.2论文内容安排.......................................................................................................4第2章人脸识别技术的综述...............................................................................................52.1人脸识别技术的基本原理............................................................................................................52.2人脸识别技术包含的内容............................................................................................................52.2.1人脸检测................................................................................................................52.2.2人脸识别...............................................................................................................72.2.3人脸图像的预处理...............................................................................................82.2.4人脸跟踪定位........................................................................错误!未定义书签。2.2.5人脸特征提取.....................................................................................................112.3基于ADABOOST算法的人脸细检测........................................................................122.3.1Adaboost算法描述..............................................................................................122.3.2Haar特征与积分图.............................................................................................142.3.3基于Haar特征的Adaboost算法描述..............................................................152.4基于隐马尔可夫模型的人脸识别方法.....................................................错误!未定义书签。2.4.1隐马尔可夫模型介绍...........................................................错误!未定义书签。2.4.2HMM的三个基本问题........................................................错误!未定义书签。2.4.3HMM人脸模型.....................................................................错误!未定义书签。2.4.4隐马尔可夫模型训练...........................................................错误!未定义书签。2.5基于嵌入式隐马尔可夫模型的人脸识别方法.......................................错误!未定义书签。2.5.1嵌入式隐马尔可夫模型.......................................................错误!未定义书签。2.5.2最佳状态链的确定...............................................................错误!未定义书签。2.5.3概率|OP的计算..............................................................错误!未定义书签。2.5.4EHMM人脸识别................................