说明：  最经典的人脸识别中的fisherface代码，在此之前要对特征空间降维，通常采用PCA降维，此代码基于降维实现类间与类内比值的最大化。
(The most classic Face Recognition fisherface code, in this feature space prior to dimensionality reduction, PCA dimensionality reduction is usually used, this code-based dimensionality reduction to achieve between-class and category to maximize the ratio.)