Matlab表情识别，特征脸[1 ]作为面部表情分类的方法。首先，利用训练图像创建低维人脸空间（pca）。这是通过训练图像集主成分分析（PCA）及图片主成分分析（即具有较大特征值的特征向量）获得的。
结果，所有的测试图像以所选择的主成分表示，计算投影图像与所有投影列车图像的欧几里得距离，选择最小值以找出与试验图像最相似的训练图像。(The feature face [1] is used as a facial expression classification method. Firstly, a low-dimensional face space (pca) is created using training images. This is obtained by training principal component analysis (PCA) of image set and principal component analysis of image (i.e. eigenvectors with larger eigenvalues).

As a result, all the test images are represented by the selected principal components, the Euclidean distance between the projected image and all the projected train images is calculated, and the minimum value is selected to find the training image most similar to the test image.)