人眼识别系统.docVIP

人眼识别系统 摘要 眼睛是人们获取外界信息最主要的途径，同时眼睛在一定程度上反映了包括个人心理意识在内的丰富信息。现在，眼睛所包含的这些有价值的信息引起越来越多人的研究，人眼检测的成果在现实生活中也得到广泛应用。如何快速准确地定位眼睛是研究者关注的重点。由于在自然条件下获得的图像不可避免含有的噪声，再加上光线、遮盖物拍摄角度等，这些为人眼检测增大了难度。 目前，人们已经提出了多种人眼检测的方法，例如对称变换法、霍夫变换法、特殊光源法、基于统计的方法、灰度投影法和基于模板的方法等，但是每个算法在具有自己优点同时又都有一定的局限性。 在人脸定位的基础上，本文使用模版匹配方法，利用模版和测试图像的相关性，判断测试图像中人眼的位置。 为了验证算法的可靠性，本文图像数据库进行测试，测试结果表明，本文算法对上述数据库中的人眼均能够实现定位，且效果良好 本文采用MATLAB语言进行仿真，实验结果达到了预定目标，证明了所研究算法的有效性。 关键词：人眼，定位，模式识别Human Eye Recognition System Abstract The eyes are the main path that most people get information from outside，meanwhile the rich information including the personal mental consciousness to a certain extent be reflected by eyes. Now，these valuable information cause more peoples research，and the eye detection results are also used widely in real life. How to locate eyes fast and accuratly and extract characteristics of the eyes discription is the focus of the researchers.For the image that obtained under natural conditions contain noise inevitably，and the influence of the light，shelter，shooting angle，all decrease the difficulty of eyes detection. At the present time，People have put forward a variety of the eye detection methods，such as symmetry transform method，Hough transform method，special light source method，method based onstatistic，grayscale projection method and the method based on template and so on. On the basis of research on human eye recognition，we presented template matching method to recognize and locate human eyes by calculating correlation index between template and test images. In order to verify the reliability of algorithm, this paper using a variety of image database for testing, test results show that this algorithm to the database are able to realize localization of human eyes, and the effect is good. This paper adopt MATLAB language to simulate，experimental result achievedintended target，the effectiveness of arithmetic investigating was to be proved. Key Words： Human eye， Recognition， Pattern Recognition. 一 绪论 1 1.1选题目的和