说明：  采用数据库为剑桥大学ORL人脸数据库，包含40个人的400张人脸图像（每人对应10张），图像为92x112灰度图像（256灰度级）对于每个人的10张图像，随机选择7张用来训练，另外3张用于测试。采用2范数最小匹配，对每个人的另外3张训练图像分别测试.
(The ORL face database of Cambridge university was adopted, containing 400 face images of 40 people (10 for each person), and the images were 92x112 gray images (256 gray level). For 10 images of each person, 7 were randomly selected for training, and the other 3 were used for testing.The other 3 training images of each person were tested with 2 norm minimum matching.)