摘要
通过对一种计算奇异类内离散度矩阵的Fisher最优判据方法的改进 ,提出一种改进的Fisher最优判据 ,并应用于人脸识别中 .在Olivetti_OracleResearchLab(ORL)和Yale标准人脸库上的识别结果显示 ,此方法比主元分析方法 (PCA)和直接线性判别分析方法 (DirectLinearDiscriminantAnalysis,DLDA)有更好、更高的识别效果 .
Fisher optimal discriminant is an efficient method in feature extraction and plays an important role in pattern recognition. A new face recognition method based on the improved optimal fisher discriminant for singular within_class scatter matrix is proposed in this paper. Comparing with PCA and DLDA methods in ORL and Yale face databases, the new method enhances the rate of recognition.
出处
《汕头大学学报(自然科学版)》
2005年第1期64-67,80,共5页
Journal of Shantou University:Natural Science Edition
关键词
人脸识别
最优
人脸库
线性判别分析
主元分析
显示
判据
奇异
矩阵
离散度
face recognition
feature extraction
fisher optimal discriminant
classifier
PCA
within-class scatter
between-class scatter