目 录
摘 要 1
ABSTRACT .................................................................................................................... 1
第一章 绪 论 ....................................................................................................................1
§1.1 研究的背景及意义...........................................................................................1
§1.2 研究的内容与发展现状...................................................................................2
§1.2.1 研究的内容与难点.................................................................................2
§1.2.2 研究的发展历程和国内外研究现状.....................................................3
§1.3 主要关键技术...................................................................................................4
§1.4 论文的组织结构...............................................................................................5
第二章 人脸检测与识别算法 ..........................................................................................6
§2.1 人脸检测的主要方法.......................................................................................6
§2.1.1 基于知识的方法(Knowledge-based methods) ...................................... 6
§2.1.2 特征不变量方法(Feature invariant approaches) ....................................6
§2.1.3 模板匹配的方法(Template matching methods) .....................................7
§2.1.4 基于表观的方法(Appearance-based methods) ......................................8
§2.2 人脸识别的主要方法.....................................................................................10
§2.2.1 人脸特征提取.......................................................................................10
§2.2.2 人脸分类方法.......................................................................................11
§2.3 人脸检测与人脸识别的异同.........................................................................13
§2.4 人脸图像数据库.............................................................................................13
§2.5 检测和识别算法性能的评价.........................................................................15
第三章 基于改进的 Adaboost 算法的人脸检测的实现 ..............................................16
§3.1 Adaboost 算法的理论基础 ............................................................................. 16
§3.1.1 集成机器学习.......................................................................................16
§3.1.2 PAC 学习模型和弱学习理论............................................................... 16
§3.1.3 Boosting 算法 ........................................................................................ 17
§3.2 Adaboost 算法及其性能分析 ......................................................................... 18
§3.2.1 Adaboost 算法 ....................................................................................... 18
§3.2.2 Adaboost 算法性能分析 ....................................................................... 20
§3.3 基于 Adaboost 的人脸检测算法研究........................................................... 22
§3.3.1 人脸特征的选取...................................................................................23
§3.3.2 特征计算...............................................................................................23
§3.3.3 构建弱分类器.......................................................................................25
§3.3.4 训练强分类器.......................................................................................25
§3.3.5 生成级联分类器...................................................................................26
§3.4 改进的 Adaboost 算法................................................................................... 28
§3.4.1 弱分类器的求取...................................................................................28
§3.4.2 基于风险敏感因子的改进...................................................................30
§3.5 改进算法的实验和结果.................................................................................33