基于matlab的人脸识别

目 录

1 引言 ............................................................................................................................................. 2

1.1 基本介绍与概念 ............................................................................................................. 2 1.2 研发历史与发展现状 ..................................................................................................... 2 1.3 研究背景与意义 ............................................................................................................. 2 2 设计方案与分析 ......................................................................................................................... 3

2.1 YCbCr空间 .................................................................................................................... 3 2.2 灰度图像转换 ................................................................................................................. 3 2.3 噪声消除 ......................................................................................................................... 4 2.4 图像填孔 ......................................................................................................................... 8 2.5 图像重构 ......................................................................................................................... 8 2.6 人脸区域确定 ................................................................................................................. 9 2.7 边缘检测 ....................................................................................................................... 11 3 程序与识别结果 ....................................................................................................................... 12 4 讨论 ........................................................................................................................................... 16 5 结论 ........................................................................................................................................... 16 致谢 ............................................................................................................................................... 17 参考文献 ....................................................................................................................................... 17

基于MATLAB的人脸初识别

王宇轩

（电子信息工程学系 指导老师：翁亚滨）

摘要：人脸识别是一门新兴的科研项目，起始于上个世纪60年代。经过几十年的发展，现已成为一项最有上升潜力的人体特征识别技术。具有广泛的应用前景，它的工作原理是借由的生物特征来确认生物个体，报告利用MATLAB软件实现人脸信息的检测与识别，在输入的整幅图像中寻找人脸区域，把图像分割成两个部分——人脸区域和非人脸区域，从而为后续的应用作准备。报告利用YCbCr空间以及二值图像实现人脸边缘分割，将真彩图像转换为YCbCr图像，根据面部肌肤在YCbCr色度空间的分布范围，设定门限阀值，实现人脸区域与非人脸区域的分割，通过膨胀腐蚀等一系列运算剔除干扰因素，再结合长宽比、目标面积等计算方法在图像中分割出人脸区域，经试验，该方法能够排除面部表情、发型、衣着背景等干扰而确定人脸区域。 关键词：matlab程序; YCbCr空间; 灰度图像; 人脸区域; 边缘检测

Abstract:Face recognition is a new research project , Begin in the 1960s,After decades of development,Now it becomes the most potential human characteristics identification technology.It has a broad prospect of application.Its working principle is to confirm individual organisms through the biological characteristics.The article uses the MATLAB software to achieve human facial information's detection and recognition,finding the human facial area in the inputed image,then devide the image into two parts,the face region and non-face region,to prepare for the subsequent application.This report uses YCbCr space and binary image to achieve human facial edge division and convert Truecolor pictures into YCbCr pictures.According to the face skin distribution in the YCbCr Chromaticity space,set threshold threshold to realize the segmentation of human face region and non-face region.Eliminate interference factors through dilation and corrosion operations,then Combine with the aspect ratio calculations and the calculation of the area to segment the face region in the image.Through the test,this method can eliminate facial expressions,hair,clothing background interference to fix the face region.

Keywords:matlab; YCbCr space; gray image; face region; edge detection

1 引言

1.1 基本介绍与概念

人脸识别就是构建一种人脸识别系统的技术，它的功能是对人脸进行定位，同时采集人脸图像，作出预处理，以确定人脸位置[1]。作为生物特征识别技术的一项发展，在计算机领域中，人脸识别是一门新兴的，借由生物特征来确认生物个体的技术。在这里生物特征以及所对应的识别方法分别是：脸部特征——人脸识别、人体掌纹、指纹——掌纹识别、指纹识别、视网膜、虹膜——视网膜识别、虹膜识别、声音频率——语音识别、目标个体体形——体形识别、形体动作幅度及力度——敲击识别、签字——签字识别等[1]。

报告中采用的人脸识别是特指对于给定的一副任意真彩图像，借由一定的图像处理及特定算法对真彩图像进行扫描从中检测出报告需要的人脸，如果是则通过对人类范围进行加框在真彩图像中标示人脸位置[2]。

1.2 研发历 …… 此处隐藏：9640字，全部文档内容请下载后查看。喜欢就下载吧 ……