基于PCA和LBP的人脸识别.pdfVIP

基于 LBP 和 PCA 的人脸识别 摘 要 在当今社会中，随着人们隐私意识的增强，对于身份识别技术也提出了更 高要求。传统的身份识别方法暴露出越来越多的缺陷，生物识别技术应运而 生。其中，由于人脸识别具有非接触性，采集简单，特征稳定等诸多优势，具 有极佳的发展前景。 本文主要讨论的是三种人脸识别算法。首先对人脸识别的研究背景，发展 现状以及研究意义做了介绍，然后分析了线性滤波，直方图均衡，同态滤波， 几何校正等多种 图像预处理方法，接着详细介绍了基于PCA 的人脸识别算法， 基于 LBP 的人脸识别算法以及基于LBP 和 PCA 的人脸识别算法，其中包括算法 的基本思想，实现的具体步骤，识别的准确率以及存在的问题，最后对此篇论 文的完成情况做出总结并指出下一步研究方向。 本文使用了 MATLAB 软件对三种算法进行了仿真，取得了良好的实验效果。 关键词：人脸识别；PCA；LBP；识别准确率 ABSTRACT In the current society, with the increase of people’s privacy awareness, they also propose higher requirements for the identity recognition technology. The traditional identity recognition method shows more and more defects. Biological recognition technology has been the tendency. Facial recognition has multiple advantages like non- contact, simple collection and stable characteristics, so it is of great development prospect. This paper mainly discusses about three facial recognition algorithm. First of all, it introduces the research background, development situation and research significance of facial recognition. Then it analyzes multiple image pre-handling methods like linear filtering, histogram equalization, homomorphic filtering and geometric correction. Then it gives a detailed introduction to the facial recognition algorithm based on PCA, based on LBP and based on LBP and PCA, including the basic thoughts of algorithm, concrete implementation steps, accuracy rate of recognition and existing problems. Finally, it sums up the completion conditions of this paper and points out the next research orientation. This paper uses MATLAB software to simulate three algorithms and gets the good experimental effects. Keywords: face recognition; PCA; LBP; accuracy rate of recognition 第一章 绪论 1.1 论文选题背景 随着时代的发展和科技