Visual C++人脸识别系统设计汇.doc

摘 要 人脸识别因为在驾驶员疲劳驾驶、视频会议、医学、档案管理、系统公安（罪犯识别等）等方面的巨大应用前景越来越成为模式识别和人工智能领域的一个研究热点，其中，由于眼睛与人脸部的其他部位相比，其灰度变化特征十分明显，因此，人眼定位研究算法对于人脸识别系统有着非常重要的意义和影响。 在常规的光线下，人眼区域在人脸区域中有两个显著地特征，一个是较周围的区域暗，另一个是灰度差值较大，因此，本文使用了一种通过灰度积分和微分投影的眼睛定位方法，其前提是在人脸区域中进行人眼定位，可用于疲劳驾驶检测中的人眼快速定位。在从人脸区域中去除容易被分隔出的，又影响人眼定位的干扰信息的基础上，利用水平和垂直灰度微积分投影的方法并结合人脸的结构特征得到眼睛的位置坐标。 该算法利用Visual C++编写程序，运算时间较短，结合简单有效的灰度投影分析，大大提高了定位的速度，并且定位较准确。 关键词：水平和垂直灰度微积分投影；灰度投影； 人眼定位  Abstract Face recognition for the pilot fatigue driving,video conference, medicine, human-computer interaction, file management, public security system (criminal identification and application prospect of the huge more and more become pattern recognition and the field of artificial intelligence of a research focus, among them, because the eye to the other parts of the face compared with the gray features very obvious, therefore, eye location algorithm research for face recognition system has a very important significance and influence. In conventional light, the human eye area in the face region are two significantly feature, a is the area surrounding the dark, the other is a gray level difference is bigger, therefore, this paper uses a through the gray integral and differential projection eyes positioning method, the premise is the face region in the eye location, can be used for the detection of the human eye fatigue driving fast positioning. From the face region in the removal of the easy to space goes out, and influence the interference of eye location information, and on the basis of using the horizontal and vertical gray calculus projection method and combining with the features of the structure of face the position of the eyes get coordinates. The algorithm using Visual C++ program, operation time is short, the combined with simple and effective gray-level projection analysis, greatly improving the positioning of the speed, and locating the more accurate. Key word:shorizontal and vertical gray calculus projection； eye location; gray integral and d