一种基于LBP特征与流形知识的人脸识别方法  

Face Recognition Based on LBP and Manifold Learning

作　　者：李菱歌 蔡超[1] 

机构地区：[1]华中科技大学自动化学院多谱信息处理技术国家级重点实验室,武汉430074

出　　处：《计算机与数字工程》2014年第7期1257-1261,共5页Computer & Digital Engineering

摘　　要：人脸识别是计算机视觉领域的研究热点,应用背景广泛。近年来,流形被认为是视觉感知的基础,流形学习算法被用来发现图像的内在特征。如何利用流形学习后的低维内蕴变量成为相关研究的核心问题。但是利用传统的流形学习算法降维得到的人脸低维特征在可分性上存在一定的不足。此外,流形学习算法对光照和姿态变化敏感。针对这两个问题,提出了一种基于局部二值模式(LBP)和流形知识的人脸识别方法。该方法首先利用LBP算子对人脸图像进行局部特征描述,然后使用流形学习算法获得高维特征数据的低维内蕴变量,并用泰勒展开式近似该流形,获取流形知识,最后利用流形知识估计流形距离来实现人脸识别。实验证明,该方法增强了人脸识别对光照变化的鲁棒性,从而提高了识别性能。Face recognition is a hot research topic in the field of computer vision ,and has wide application .Recently manifolds are thought to be fundamental for visual perception ,and manifold learning algorithms are developed for discovering intrinsical features .How to use the low dimensional intrinsic variables obtained by manifold learning becomes the core issue of related research .But the classification result sometimes is not accurate when faces are classified in low dimensional sub-space directly .Furthermore ,manifold learning methods are sensitive to the variation of illumination conditions .In order to solve these two problems ,a novel face recognition method based on LBP and manifold learning is proposed .Firstly ,LBP op-erator is used to describe the local features of the face images .After obtaining the intrinsic variables of the feature data by manifold learning algorithm ,the manifold is then approximated by higher-order Taylor expansion whose parameters are saved as manifold knowledge .And then face recognition is realized by solving the manifold distance .Experimental results show that the proposed method is robust to illumination and can improve face recognition performance effectively .

关 键 词：人脸识别 局部二值模式(LBP) 流形学习 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]