模糊二维线性鉴别分析算法

摘要

This paper proposes a new method of face image feature extraction,namely,the fuzzy 2DLDA based on the 2D Linear discriminant criterion and fuzzy set theory.In the proposed method,we calculate membership degree matrix by FKNN,then we incorporate the membership degree into the definition of the intra-class scatter matrix and inter-class scatter matrix and get the fuzzy intra-class scatter matrix and fuzzy inter-class scatter matrix.Experiments on the Yale,ORL and FERET face databases show that the new method can work well.%在二维线性鉴别分析的基础上,引入了模糊数学的思想,给出了模糊二维线性鉴别分析方法。模糊二维线性鉴别分析通过隶属度函数将样本归入所有的类别之中,根据隶属度重新定义了类间散布矩阵和类内散布矩阵,进而将样本的原始分布信息通过相应的隶属度函数完全融入到最后提取的特征中。通过在ORL和耶鲁人脸库上进行的实验,结果证明了基于模糊二维线性鉴别分析方法的优越性。