摘要: 伴随着互联网逐渐走进人们生活，人们对信息安全看的越来越重要。人脸识别相比较于指纹、手型、虹膜、语音等生物特征来说，前者拥有快速、友善、稳定的特点，更容易被使用者接受，同时该项技术普遍应用于医疗研究、身份验证、公安鉴别、视频会议等领域。42111
本文从研究与仿真两方面对课题进行探讨，首先，采用中星微芯片USB摄像头采集人脸信息；然后进行图像预处理，包括滤波去噪、图像灰度化、直方图均衡化等，基于Adaboost算法的人脸检测定位，采用Haar-Like进行特征提取；最后结合传统PCA算法以及改进后的PCA+LDA算法对人脸库进行识别，通过Matlab仿真实验比较二者的性能，得出后者在相同环境下识别率更高更稳定的结论。42111
毕业论文关键词: PCA ; PCA+LDA; 人脸识别; MATLAB
Face Recognition System Design
Abstract: With the Internet gradually come into people's lives, people look more and more important to the security of information. Compared to the face recognition compared with fingerprint, retina, iris, gene biometrics, face recognition has characteristics of fast, friendly, stable, and thus more likely to be acceptable to users, also face recognition technology widely used in medical research, authentication, public security identification, video conferencing and other fields.
From the research and Simulation of facing the topic to discuss, first of all, using Vimicro USB camera to capture the face information and of face image pre processing include filtering noise, the gray image, histogram equalization, etc., uses the method based on the explicit feature of face detection and location, using Haar like feature extraction, finally, it combines with the traditional PCA algorithm and improved PCA + LDA Algorithm for face database to identify, through the MATLAB simulation experiments compared the performance, finally comes to the conclusion that the latter in the same environment recognition rate taller and more stable.
Key words: PCA; PCA + LDA; Face Recognition; MATLAB
目 录
摘要 2
引言 3
1. 课题研究现状及面临问题 4
1.2 当前课题存在的主要问题 5
2. 人脸识别系统 5
2.1 人脸识别研究的目的和意义 6
2.2 人脸图像采集 7
2.3 图像预处理 8
2.3.1 滤波去噪 8
2.3.2 图像灰度化 8
2.3.3 直方图均衡化 9
2.4 人脸检测 9
2.4.1 基于Adaboost算法 10
2.5 特征提取 11
3. 基于PCA+LDA算法的人脸识别 11
3.1 PCA 算法分析 11
3.2 LDA算法 12
3.3 PCA与LDA结合的人脸识别 13
3.4 本章小结 14
4. 基于Matlab的仿真分析 14
4.1 Matlab简介 14
4.2 系统软硬件平台 14
4.3 系统执行界面 Adaboost+Haar-Like人脸识别系统的设计+matlab代码:http://www.youerw.com/zidonghua/lunwen_42440.html