Matlab实时监控系统的人脸识别系统设计+文献综述
在本文中，我们重点研究多种不同的人脸特征和技术在变化的光照条件下的识别性能。本文根据Matlab计算软件，分别采用DCT离散余弦变换技术，LBP以及LLDP技术来进行光照不变的人脸识别，其中基于DCT的方法主要思想是丢弃一些低频系数来减少光照变化对人脸特征的影响。而LBP，LLDP局部特征方法是将图像分割若干块，并在每个块内将其中每一个点与周围的点比较，最后整体比较图像的直方图来判断人脸相似度。实验证明这三种技术都能有效的提高人脸识别系统在光照变化条件下的识别性能。27269
关键词 人脸识别；光照变化；离散余弦变换(DCT)；局部二值化模式(LBP)；局部线形导数模式(LLDP)。
毕业论文设计说明书外文摘要
Title Face recognition system in real-time surveillance
Abstract
Face recognition is one of the most popular biometric modalities based on human facial feature. Because of its user-friendly, fast and convenient advantages，it is easy to be accepted by users, and therefore has been widely studied and has a lot of applications. Existing face recognition system have achieved satisfactory performance under controlled environment. However in the real applications, due to uncontrolled variations such as illumination, pose, expression and aging, the performance of existing system are still unsatisfactory.
In this paper, we mainly focus on studying recognition performances of several facial features and technologies under varying illumination conditions. Based on the MATLAB, we apply Discrete Cosine Transform (DCT), local binary pattern (LBP) and local line derivative pattern (LLDP) respectively in illumination invariant face recognition. The main idea in the method based on the DCT is to discard some low-frequency coefficients to remove the effect caused by illumination variations, while the methods based on local features such as the LBP, LLDP are to pide the image into several blocks, compare each point with the neighbor points in each block, and calculate the histogram of local features to measure the similarity of facial images. Experimental results demonstrate that all of the above three techniques can improve the performance of face recognition system under varying illumination conditions.
Keywords Face recognition；Illumination Variations；Discrete Cosine Transform ； Local Binary Pattern； Local Line Derivative Pattern.
目 次
1引言 1
1.1课题背景… 1
1.2人脸识别研究意义…1
2 文献综述… 3
2.1常见人脸识别系统…3
2.2 人脸识别主要挑战 4
2.3常见的光照不变人脸识别技术5
3 基于DCT的光照不变人脸识别方法… 8
3.1介绍…8
3.2 方法及其原理 8
3.3实验结果分析 10
4 基于LBP的光照不变人脸识别方法…14
4.1介绍 14
4.2方法及其原理 14
4.3实验结果分析 15
5 基于LLDP特征的光照不变人脸识别方法 18
5.1介绍 18
5.2方法及其原理 18
5.3实验结果分析 20
结论23
致谢 … 24
参考文献25
1 引言
随着科技的发展，人们对于安全的重视以及金融贸易等各种应用的需求，生物识别技术得到新的重视，而人脸识别技术是最近几年兴起的生物识别技术方法中应用最广泛的技术之一。