基于彩色信息融合和同伦算法的遮挡鲁棒人脸识别方法研究  

Study of Occluded Robust Face Recognition Approach Based On Homotopy Algorithm and Color Information Fusion

作　　者：翟懿奎[1,2] 甘俊英[2] 李景文[1] 

机构地区：[1]北京航空航天大学电子信息工程学院,北京100191 [2]五邑大学信息工程学院,江门广东529020

出　　处：《信号处理》2011年第11期1762-1768,共7页Journal of Signal Processing

基　　金：国家自然科学基金项目(No.61072127);广东省自然科学基金项目(No.10152902001000002;No.S2011010001085;No.S2011040004211;No.07010869);广东省高等学校高层次人才项目(粤教师函[2010]79号);广东高校优秀青年创新人才培养计划项目(粤财教[2008]342号);广东省产学研合作项目(NO.2009B090300416);五邑大学青年科研基金(Q948)

摘　　要：遮挡条件下的鲁棒人脸识别,目前在人脸识别领域逐渐被重视,被认为是难点问题之一.本文利用稀疏表示理论满足人眼视觉特性及神经信息有效表达,且跟人脸固有特征具有的自然性是相吻合的特点,研究了彩色人脸图像色度信息有效融合策略,采用同伦算法解决稀疏表示模型中的I_1范数问题,提出了一种基于彩色信息融合和同伦算法的遮挡鲁棒人脸识别算法.在AR数据库中的实验结果表明,与传统基于灰度转换方法人脸识别方法及SRC算法相比,本文所提基于同伦算法的稀疏表示人脸识别,具有很高的计算效率,而且有效融合了彩色信息,显著提高了在遮挡及非遮挡情况下人脸识别的效率及识别性能。Robust face recognition（FR） under occluded condition is considered more and more important gradually in FR field, and it is one of the difficult problems.While sparse representation theory hit the spot of human visual characteristic and neural information effective expression,and it＇ s consistent with the human face inner feature,the effective fusion strategy of color face information is studied in this paper,and the homotopy algorithm is used to solve the I_1 norm problem in occluded sparse representation based FR.Experimental results in AR database show that compared with traditional color face image fusion FR method and the traditional gray-scale conversion FR method,the homotopy algorithm and color information fusion based FR method can achieve high recognition performance in both unoccluded and occluded face image.It is also show that the effective integration of color information on feature fusion,also contributes to improve the efficiency and performance of face recognition system.

关 键 词：人脸识别 彩色信息融合 稀疏表示 同伦算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]