Adaboost算法基于人脸特征的性别识别

摘要人脸是重要的生物特征之一，人脸图像上蕴含了大量的信息。随着技术的进步，基于人脸图像的模式识别问题，近年来已成为研究的热点。

我们知道让计算机根据输入的人脸图像判断性别并不容易。本文基于人脸特征进行性别分类，总结了基于知识和基于统计理论的人脸检测方法，并着重介绍了Adaboost算法的基本原理。为了提高性别识别率，文中在比较多种检测方法优缺点的基础上，提出了采用Adaboost算法对人脸进行检测。66379

在特征提取方面，本文概述了压缩感知的基本理论，用基于压缩感知的方法对人脸进行特征提取，为接下来的实时性别识别奠定了良好的基础。

支持向量机(SVM)在解决小样本问题以及高维问题有比较明显的优势，并且具有一定推广能力。文中使用支持向量机对性别分类。实验选取径向基核函数，取得了不错的性别识别率效果。

毕业论文关键词：人脸检测压缩感知支持向量机性别识别

毕业设计说明书（论文）外文摘要

Title Gender based on facial recognition feature

Abstract

Face is an important feature of the biological，Face images contain a great deal of information.With advances in technology，face image based pattern recognition problems，recent years have become a hot research.

We know that let the computer according to the input face image is not easy to determine the sex.This article based on facial features for gender classification，summarizes the statistical theory based on knowledge and based on face detection method，and highlights the basic principles of Adaboost algorithm.In order to improve gender recognition rate, the paper comparing the advantages and disadvantages of various detection methods, based on the proposed use of Adaboost algorithm for face detection.

In the feature extraction, this paper outlines the basic theory of compressed sensing,take methods based on compressed sensing , for the next time gender recognition has laid a good foundation.

Support vector machine (SVM) in solving small sample size problem and high dimensional problems are more obvious advantages, and has a certain ability to promote.Use the support vector machine for sex.Experimental select RBF kernel function, and achieved good gender recognition rate results.

Key Words: Face Detection compressed sensing support vector machine gender recognition

第一章绪论 1

1.1 性别识别问题概述 1

1.2 性别识别的研究意义及典型应用 1

1.3 性别识别的研究现状 2

1.4 论文组织结构 3

第二章人脸检测 4

2.1 人脸检测方法 4

2.2 人脸图像库 4

2.3 基于AdaBoost算法的人脸检测 5

2.4 本章小结 11

第三章基于压缩感知的人脸特征提取 12

3.1 压缩感知理论介绍 12

3.2 基于压缩感知的人脸特征提取 15

3.3 特征提取过程与结果 17

3.4 本章小结 18

第四章性别识别 Adaboost算法基于人脸特征的性别识别:http://www.751com.cn/zidonghua/lunwen_74305.html