主分量分析（PCA）和线性鉴别分析（LDA）是模式识别领域的使用最为广泛的两种特征抽取方法，而在图像识别中经常采用的是PCA+LDA方法来代替单纯的LDA。本文提出一种增强型线性鉴别准则（ELDA），将PCA的优点和LDA的优点充分地融合在一起，不仅解决了PCA过程中使用最小距离方法时识别精度相对低的缺点，而且解决了LDA过程中当类内散布矩阵奇异时投影向量的求解问题，也就是说可以使用该方法来替代PCA+LDA的两步骤方法。另外，该方法在识别精度上比PCA和LDA或PCA+LDA方法都有较大的提高，通过在ORL、Yale和NUST603人脸库上的实验验证了该算法的有效性。

Principal Components Analysis (PCA）and Linear Discriminant Analysis (LDA）are two popular feature extraction methods for pattern recognition,and in image recognition,researchers usually use PCA+LDA instead of LDA.An enhanced linear discriminant analysis (ELDA）criterion,which integrates their merits,is proposed in the paper.It can not only overcome the PCA’s shortcomings of lower precision when using the minimal distance,but also resolve the problem of projective vector solution of LDA when the within class scatter matrix is singular.So the two step method of PCA+LDA can be substituted by ELDA.Moreover,its recognition rate exceeds the single PCA,LDA,or PCA+LDA largely.Many experiments on ORL,YALE and NUST603 face database indicate that our method is effective.