【作者】 叶龙宝；

【导师】 高建生；

【作者基本信息】 华中科技大学 ， 软件工程， 2011， 硕士

【摘要】 随着社会的不断进步,科学技术的飞速发展,计算机应用领域也在不断拓展,于是数字图像处理技术作为一种全新的图像处理方法应运而生,而且应用范围越来越广。数字图像处理技术是利用计算机设备将图像转变成数字信息来进行保存、处理、传输和重现,通过计算机对图像进行去除噪声、增强、复原、分割、提取特征等处理的方法和技术。人脸识别技术是计算机模式识别技术在数字图像领域中的具体应用,广泛应用于安全验证系统、档案管理、视频会议、人机交互等方面,因而越来越成为当前模式识别和数字图像处理领域的一个非常热门的研究课题。目前,数字图像处理与人脸识别技术相结合的应用范围越来越广,已经深入到人们的生活之中。但就目前计算机的应用水平而言,由于计算机对外部的感知能力还比较弱,自动化的处理能力还有待提高,仍然需要投入大量人力、物力来从事数字图像处理与人脸识别相结合的理论和应用的研究。本课题论述了人脸识别技术在图像处理中的一个应用,采用几何特征的方法,包括人脸图像的预处理、人脸轮廓的检测、人脸图像中眉毛、眼睛、鼻子、嘴巴各部位的自动定位以及人脸图像的识别等几大功能。其中人脸检测部分,采用了基于haar-like特征的人脸检测算法,该算法在具有较高的检测率的同时也满足了实时检测的要求。算法利用OpenCV提供的库结合层叠式分类器并使用一系列haar-like特征来描述人脸。在人脸面部特征定位的研究中,采用了基于主动形状模型(ASM)的方法,能够很快的定位到人脸面部的各个特征部位。本系统的出现为人脸识别技术在数字图像处理方面的应用寻找到了一个新的结合点,也使数字图像处理技术的应用范围更加广泛。更多还原

【Abstract】 With the double-quick development of society and science technology, the field in computer application is also expanding as soon as possible.So digital image processing techonology has emerged as the times require as a bran-new image processing method.As the same time, it is also applied to more and more domains. Digital image processing is a method and technology which uses computer devices to convert image to digital informations in order to save,process,transmit and recur,and process image for wipe off yawp,enhance,recove,partition and extraction.Face recognition technology as a particular application of pattern recognition in image field is widely applied to surveillance and security, digital libraries,video meeting,and human-computer intelligent interaction and so on.The application area combining digital image process with face recognition technology is more and more extensive and embedded to people’s life at present. But as the present computer level, the ability of computer in outside apperception is still very weak, so it needs more manpower and more material resource to do research on the theories and applications of combining digital image processing with face recognition. In this paper, we discuss one of the applications of face recognition among image processing which is based on geometry feature method. This system includes several functions such as pre-processing of facial images、face figure diction、automatic location of the eyebrow、eye、nose and mouth,and recognition of face images.In the part of face diction, the method used in this thesis is face diction algorithm based on haar-like feature. It can provide with high accurate, meet the request for real time detection, use the library provided by OpenCV combining with cascade classifier to describe the face using a serial of haar-like features. In our researehes for facial features localization,the method of active shape model(ASM) was used in order to locate each characteristic part of face quickly.This system will offer a new combinable point, which is that face recognition technology is applied in digital image processing, and expand the range of digital image processing application.更多还原