论文总字数：26703字

摘 要

随着科技的进步、社会的发展，人类对于身份识别有了更加深刻的认识。在诸多身份识别的技术中人脸识别技术用于个人身份鉴定有一定的优势。人脸识别技术使用较为方便，基本上只需要摄像头就能采集基本的信息，这种无需直接接触的方式简单方便容易被人接受。人脸识别技术不易被人仿冒，而且相对于其他的身份识别技术有高精确度、低错误识别率和较低的拒认率等优点。人脸图像的采集设备非常普遍，常见的PC、摄像机等通用设备就能实现人脸识别技术的基本功能。对大多数应用而言无需添加任何大型设备就能使用人脸识别功能，既方便又安全。

人脸定位系统基本都包括图像采集、图像预处理、特征提取、特征匹配、识别结果。

人脸定位系统的实现是通过对预处理后的图像进行边缘检测，利用人眼在脸部的位置特性和人眼区域特征再结合投影法找出人眼的区域。具体的做法是先对人脸图像进行边缘检测提取人脸边缘，通过人脸边缘的提取凸显人的眼睛、鼻子、嘴巴的位置，结合投影法获得人眼的大概范围，通过阈值法对该范围进行筛选合并确定人眼的位置。在眼睛已经定位的情况下对鼻子定位：依据双眼的距离和平均高度计算鼻子的上下边界和水平中点，根据双眼的距离和斜角及鼻子的中点计算左右边界。再对嘴巴进行定位：双眼距离、高度、斜角可以找到嘴巴的上下边界和水平中点，再利用肤色掩码可以定位嘴巴。最终成功完成了人脸部五官的基本定位。

关键词：人脸定位 边缘检测 投影法

The Realization of Face Location System

Abstract

Along with the progress of science and technology and the development of society, people have a more profound understanding of the identification of identity. Face technology is more convenient to use, basically just need a camera to collect the basic information, this kind of non-contact way is simple and easy to be accepted by people.face recognition technology is not easy to be fake, and relative to other identity recognition technology has the advantages of high precision, false recognition rate and lower rejection rate. Face image acquisition device is very common, common PC, camera and other common equipment can achieve the basic function of face recognition technology. For most applications, the face recognition function can be used without adding any large equipment, which is convenient and safe.

Face positioning system includes image acquisition, image preprocessing, feature extraction, feature matching, recognition results.

Face positioning system is realized by the image edge detection, the use of the human eye in the face of the location and the characteristics of the human eye region combined with projection method to identify the region of the human eye. Specific approach is to first on the face image using multi threshold binarization, the eyes in the face of the above, look for the area the biggest connected area (hair); combined with the projection method to obtain human about the scope of, by threshold the range of screening combined to determine the location of the human eye. Under the eyes is the positioning of the positioning of the nose: Based on binocular distance and average height calculation of the nose under the boundary of the level and the middle. According to the midpoint of the binocular distance and angle and nose count of the left and right boundaries. Re positioning of the mouth: binocular distance, height, angle can be found in the upper and lower boundaries of the mouth and the level of the point, and then use the color mask can locate the mouth. In the end, the eyes, nose and mouth of the human face are successfully realized.

Key Words: Face location edge detection projection

目录

目录

摘 要 I

Abstract II

第一章 绪论 1

1.1 人脸定位的选题背景和研究意义 1

1.2 人脸识别的发展 1

1.3课题的主要工作和结构安排 2

第二章 人脸识别系统的概述 3

2.1人脸识别系统的基本构成 3

2.2人脸识别系统的各个步骤 3

2.2.1图像采集 3

2.2.2图像预处理 3

2.2.3特征提取 4

2.2.3特征匹配 4

2.2.4识别结果 4

第三章 人脸区域的检测 5

3.1 色彩模型 5

3.2肤色建模和肤色区域 7

3.3人脸区域分割 7

第四章 人脸部器官定位方法的实现 10

4.1 脸部器官定位的整体设计 10

4.1.1人脸面部的几何关系 10

4.1.2人脸器官定位过程 10

4.2人脸边缘提取与检测 11

4.2.1边缘检测 11

4.3 眼睛的定位 14

4.3.1眼睛定位方法概述 14

4.2.2本文的方法 15

4.3鼻子的定位 23

4.4嘴巴定位 26

第五章 研究结果 29

5.1 测试过程与结果 29

5.2 性能分析 31

第六章 总结与展望 32

6.1 总结 32

6.2 展望 32

参考文献 34

致谢 36

请支付后下载全文，论文总字数：26703字