说明：  1. 本程序使用PCA算法对训练图像降维，得到特征图像； 2. 将每一幅图片与平均图像的差值投影到特征向量空间； 3. 计算训练图片投影两两之间的最大欧氏距离distance_max； 4. 将测试图片也投影到特征向量空间，计算它与所有训练图像投影的最小欧氏距离distanceST_min。 5. 设定判别阈值（设为0.3*distance_max，此为测试得到的经验值，不同阈值将影响判别结果）。当distanceST_min>0.3*distance_max时，则认为测试图像不属于和训练图像相同的类，反之，则认为是。
(1 This program uses PCA dimensionality reduction algorithm training images obtained feature image 2 would mean every picture with the image projected onto the difference feature vector space 3. Computing training images projected maximum pairwise Euclidean from distance_max 4. the test image is projected to the feature vector space, with all the training images calculate its minimum Euclidean distance projection distanceST_min. 5 Set judging threshold (set 0.3* distance_max, the experience obtained by the test, different threshold values ​ ​ will affect the determination result). When distanceST_min> 0.3* distance_max, then that test image and the training images do not belong to the same class, on the contrary, it is considered to be.)