结合奇异值和形态学的人脸特征提取  

Face Feature Extraction Based on Singular Values and Morphology

作　　者：刘艳艳[1] 安霞[1] 

机构地区：[1]防灾科技学院基础课教学部,三河065201

出　　处：《现代计算机（中旬刊）》2013年第8期16-19,共4页Modern Computer

基　　金：中国地震局教师科研基金项目(No.20110116);河北省自然科学基金项目(No.A2011408006);中央高校基本科研业务费专项资金青年教师计划项目(No.ZY20110225)

摘　　要：奇异值特征包含的鉴别信息不足以正确识别人脸等相似对象,针对此问题提出一类用于人脸识别的增强奇异值特征。利用数学方法将包含较多局部鉴别信息的区域与全局图像融合,得到与奇异值特征具有类似属性的新特征,在人脸识别实验中表现出良好的识别分类能力。此外提出一种基于灰度形态学的五官图像提取方法,对人脸的姿态、表情等变化不敏感,实验表明其满足增强奇异值特征对局部区域的提取需求。As a kind of algebraic features to describe face images, singular values neglect the character of human faces and contain inadequate information for recognition. To address these problems, proposes a new kind of features named as enhanced singular values, which is based on the global and local information. Enhanced singular values utilize the high discriminatory local regions to recognition, and have the beneficial properties similar to singular values＇. The experimental results on face databases indicate the new features are effective for recognition. Furthermore, proposes a new robust method based on gray morphology to detect the needed local regions. The experimental results illustrate is robust for the variations of facial details, pose and expression, and appropriate for extraction of enhanced singular values.

关 键 词：特征提取 奇异值分解 灰度形态学 局部区域 人脸识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]