Face Recognition Based on Covering Algorithm and PLS

This paper proposes a new approach of human face recognition using partial least square (PLS) and covering algorithm. PLS constructs the component so that the correlation between the class variable and themselves maximized. Covering algorithm is a kind of neural network method. The approach first processes the image with wavelet transform, and then eliminates the redundant features with PLS. It reduces the dimension of image by extracting the principal features. Finally, classifies the image by covering algorithm