毕业论文--人脸检测系统设计

《毕业论文--人脸检测系统设计》由会员分享，可在线阅读，更多相关《毕业论文--人脸检测系统设计（105页珍藏版）》请在毕设资料网上搜索。

1、 中文题目：人脸检测系统设计 外文题目：FACE DETECTION SYSTEM 毕业设计（论文）共 100 页（其中：外文文献及译文 14 页）图纸共 0 张 完成日期 2015 年 6 月 答辩日期 2015 年 6 月 I 摘要 本项目致力于完成 Microsoft Visual Studio 2010 平台下的人脸检测与定位系统， 人脸检 测定位的算法基于基于肤色相似度算法。本设计主要完成的工作包括求肤色的相似概率、 二值化、水平与垂直方向的投影、人脸特征（眼睛、眉毛、嘴巴和鼻子）的提取、精确定 位人脸区域。系统对于提升人脸检测的准确率具有指导意义， 设计了基于肤色的人脸检测和定位系

2、统。 采用了脸部皮肤之间相似度的方法以及二 值化方法，使用了基于边界方法和基于区域方法相结合的算法， 提取了眼睛、嘴和鼻子 等关键特征，最终较好地实现了人脸定位。 在 Microsoft Windows 平台上，利用 Microsoft Visual Studio 2010 开发了软件。 对于清晰的正面人脸图，该系统能够做出正确地检测定位并提取出特征。 关键词：图像分割；人脸定位；肤色；人脸检测；特征提取 II Abstract This topic devotes to completing the detection and localization system of the perso

3、n face under the Microsoft Visual Studio 2010 platforms, and the detection localization algorithm of the person face limit to basing on the skin color similarity algorithm. The topic must complete similarity computation, binary image processing, vertical histogram, horizontal histogram and extractin

4、g person face characteristic (eye, mouth and nose). This topic emphasize application of skin color model similarity in the detection and the localization of person face, and has a significant instruction for research of increasing accuracy in detection and the localization of person face. In this pa

5、per, the authors have presented an algorithm and designed a system for face detection and location based on complexion. By strengthening the contrast between face features and by adopting binary image processing method, the system has improved the preprocessing effect; and by using boundary-based al

6、gorithm plus region-based algorithm , the system has realized face location through the extraction of the features of eyes, nose and mouth. Taking advantages of Microsoft Visual Studio 2010, the authors have also developed corresponding software based on Microsoft Windows. Production of this paper have definite application value.Experiment results prove that the system is valid in detecting. Key words: image segmentation; face localization; complexion; face detection; feature ex-traction