matlab编写的Modular PCA的源代码，以Yale人脸库为例。Moduplar PCA首先对原始图像分块，然后对分块后的所有子图像进行PCA分析提取投影特征，对待识别图像也是先进行分块，然后分别计算子图像在投影特征下的投影系数，最后根据最近邻分类器进行分类。-Modular PCA prepared matlab source code to Yale face database as an example. Moduplar PCA first block of the original image, and then on the block after all the sub-image analysis to extract PCA projection characteristics, treatment identification image is first block, and then calculated the characteristics of sub-image in the projector under the projection coefficient, and finally the light of recent Neighbor Classifier classification.