文档介绍：提要
人脸识别技术是基于生物特征的识别方式,与指纹识别等传统的识别方式相比,具有实时、准确和非侵扰等特性。因此人脸识别技术在诸多领域都有广泛的应用。人脸识别技术中的人脸特征提取及模式识别是近年来基于生物特征研究的热点之一。
本文以人脸识别的关键步骤为主要内容,系统地研究了人脸识别技术,对各环节所需的算法作了介绍和研究。在预处理环节中,对图像进行大小归一化及灰度归一化等处理。在特征提取环节,介绍几种特征提取的方法并对其进行对比,深入研究基于主成分分析(PCA)的特征子空间方法提取本征脸(Eigenface)。在分类器识别环节,对非线性Parzen分类器中核函数的参数估计算法和欧氏距离分类器法进行比较,重点讨论采用欧氏距离分类器的识别法。
关键字
预处理;PCA;欧氏距离分类器;人脸识别
Face Recognition Based on PCA
Research and Implementation
060608117 Lin Xiaoming Tutor:Chen Yu Lecturer
Abstract
Face recognition technology is based on biometric identification methods such as fingerprint recognition and identification of the traditional methods, with real-time, accurate and non-intrusive. Face Recognition Technology Face feature extraction and pattern recognition in recent years based on the biological characteristics of one of the hot spots.
In this paper, a key step in face recognition as its main content, a systematic study of the face recognition technology,required on the part of the algorithm was introduced and research. In the preprocessing stage,In the pretreatment session, the image size normalization and gray normalized such proceedings. In the feature extraction part, introduces several feature extraction methods and contrast,extracting Eigenface Based on ponent analysis(PCA) of the subspace was studied. Identify areas in the classifier, Parzen classifier in the nonlinear kernel function parameter estimation algorithm and Euclidean distance classifier method parison, focused on the use of Euclidean distance classifier recognition method.
Keywords
Preprocessing ; PCA ; Euclidean distance classifier ; Face recognition
目录
第一章绪论 - 1 -
研究的背景与意义 - 1 -
人脸识别的国内外研究现状 - 2 -
人脸识别在国外研究现状 - 2 -
人脸识别在国内研究现状 - 5 -
课程研究的内容 - 6 -
第二章人脸图像的预处理 - 8 -
第三章人脸的特征提取 - 11 -
ICA及其人脸表征 - 12 -
ICA的基本思想 - 12 -
人脸的独立分量表征 - 12 -
LDA 算法及其人脸表征 - 13 -
LDA 算法 - 13 -
LDA表征人脸 - 14 -
PCA特征提取方法 - 15 -
K-L变换的基本原理 - 15 -
PCA基本原理 - 17 -