摘要： 为提高对彩色人脸的识别率，提出一种基于Fisher块对角局部非负矩阵分解(LNMF)的识别算法。采用块对角矩阵编码彩色图像不同通道的颜色信息，在LNMF算法中增加块对角约束和Fisher判别约束，对不同通道的颜色信息同时进行计算并融入人脸的类别信息，用于提取人脸特征。在CVL和PIE彩色人脸数据库上的实验结果验证了该识别算法的有效性。

关键词: 人脸识别, 非负矩阵分解, Fisher判别

Abstract: To improve the recognition rate of color face, this paper proposes a recognition algorithm based on Fisher block diagonal Local Non-negative Matrix Factorization(LNMF). The algorithm employs block diagonal matrix to encode color information of different channels for color images. Block diagonal constraint and Fisher discrimination constraint are imposed on LNMF algorithm to compute color information of different channels simultaneously for face feature extraction. Experimental results on CVL and PIE color face databases verify the effectiveness of the recognition algorithm.

Key words: face recognition, Non-negative Matrix Factorization(NMF), Fisher discrimination

中图分类号: 

• TP391.41