摘 要Abstract 人脸检测的研究具有很重要的意义,可以应用到人脸识别、新一代的人机界面、安全访问和视觉监控以及基于内容的检索等领域,近年来受到研究者的普遍重视。人脸检测的目的是要从图像中把人脸分割出来。 The person face examination's research has the very vital significance, may apply the person face recognition, new generation's man-machine contact surface, safe visit and visual monitoring as well as based on content domains and so on retrieval, in recent years received the researcher universal recognition. The person face examination's goal is must divide from the image the person face. AdaBoost算法是1995年提出的一种快速人脸检测算法,它的出现,大大提高了人脸检测的速度和效率,是人脸检测领域里程碑式的进步。本文前两章描述了人脸检测的一般方法和历史概况。 The AdaBoost algorithm is one kind of fast person face examination algorithm which in 1995 proposed, its appearance, raised the person face examination speed and the efficiency greatly, is the person face examination domain milestone-like progress. This article first two chapters described the person face examination general method and the historical survey. 第三章讲述了AdaBoost算法的历史和数学模型,多级分类器的构建过程。 第四章给出了本文的实现方法,并给出了实验结果,与Viola作了对比,并在附录中给出了程序 The third chapter narrated the AdaBoost algorithm history and the mathematical model, multistage sorter's construction processes. the fourth chapter gave this article to realize the method, and has given the experimental result, has made the contrast with Viola, and has given procedure in the appendix 关键词:Key word: 人脸检测Person face examination 分类器Sorter haar特征haar characteristic