毕业论文定稿基于Android的移动人脸识别系统设计：人脸预处理与辨识精讲.doc

摘要 人脸识别（Face recognition）是基于人的脸部特征信息身份识别的一种生物识别技术 Abstract Face recognition is a kind of biometric identification technology based on facial features of peoples information, is a research subject of artificial intelligence in the field of biological identification is very important and practical. Compared to the fingerprint and iris recognition, face recognition is non-contact, use more intuitive and convenient. In recent years, face recognition technology has attracted more and more attentions, progress has been made largely of face recognition, it is now widely used in various fields of daily life, such as network security, public security, electronic commerce, finance, attendance and property management etc. In this paper, the aspects of face preprocessing, feature extraction and classification and recognition are studied and analyzed.?Firstly, the influence of the factors such as ray, attitude, expression and angle on the image quality is analyzed, and the necessity of face preprocessing is clarified, and the common preprocessing method is introduced. Then, the Gabor wavelet transform and principal component analysis (PCA) are introduced and analyzed, which is used for the feature extraction of human face.?In this paper, the feature of face is characterized by the nearest neighbor classifier and SVM classifier. And experiments were conducted using the AR face database, respectively by different experimental results to analyze the performance of pre treatment, Gabor wavelet and PCA, and the nearest neighbor and support vector machine classifier performance comparison. 【Key Words】 Face recognition Preprocessing Gabor wavelet Support vector machine  目录 前言 1 1 人脸识别介绍 2 1.1 人脸识别发展历史 2 1.2 人脸识别系统构成 2 1.2.1 人脸图像采集及检测 2 1.2.2 人脸图像预处理 3 1.2.3 人脸图像特征提取 3 1.2.4 人脸图像匹配与识别 3 1.3 人脸识别常用方法 3 2 基于人脸的预处理 5 2.1 直方图均衡化 5 2.1.1 基本原理 5 2.1.2 实现方法 6 3 人脸特征提取与识别 9 3.1 Gabor小波变换基本原理 9 3.2 主成分分析法基本原理 9 3.3 分类方法 10 3.3.1 最近邻法（NN）基本原理 10 3.3.2 支持向量机（SVM）基本原理 12 4