文档介绍：摘要随着信息化领域的扩大,人机交互成为信息化社会的发展趋势,是未来实现人类与智能计算机便捷交流的一种有效手段。人机交互推动了生物特征识别技术的发展,人脸识别凭借其使用简单、方便、直接和人性化的特点,成为了研究热点,受到了越来越多的关注。近年来,基于纹理特征分析方法的人脸识别取得了一定的成功。本课题根据要求,采用改进的 LBP 特征的方法实现了一个实时的人脸检测和识别系统。首先, 本文从纹理特征出发,对基于 LBP 特征的人脸识别进行了分析和研究;然后,针对 LBP 特征在人脸识别中存在的缺点,提出了对脸部不同区域进行加权操作的方案,对每个区域进行分析时,为了对高维数据进行降维,采用了 Fisherface 方法, 取得了良好的效果;最后,通过 VC 程序编写软件实现算法,软件系统依靠 USB 摄像头采集视频图像序列,进行实时人脸检测,并对检测出的人脸进行拍照,组***脸库进行训练,最终实现自动人脸识别功能。经测试,采用 LBP 特征方法实现的自动人脸识别系统,有效克服了光照变化带来的影响,准确率和识别率达到了课题要求。关键词:人脸识别人机交互 LBP 特征 Fisherface 方法 Abstract With the expansion of the information fi eld, puter interaction es the trend of the information society. It will be an effective mean s to realize convenient communication between human and intel puter in the future. The puter interacti on boosts the development of biometric identification technology. Face recognition has e an im portant research point and gain widely attention by virtue of simplicity, conv enience, directness and humanization. In recent years, face recognition based on the method of textur e feature analysis has gained some ess. According to the re quirements of this research topic, this paper uses an improved method of LBP featur es and realizes a system that can do real-time face detecti on and recognition. Firstly, in this paper, the face recognition based on the LBP features space is studied and anal yzed from the texture features of images. Then, this paper presents a scheme that adds weights to different area of the face in view of the disadvantages in face rec ognition using LBP features. In addition, for dimensionality reduction, the author uses th e method of Fisherface. Finally, the software is implemented on Visual C++ platform. This system captures video images from USB camera, transmits them puters, and detects the face in real time. The faces detected are caught and make up the image da tabase. They will be trained and realize the function of automatic face recognition. After series of tests, the automatic f