Pixel selection based on discriminant features with application to face recognition

We propose a pixel selection method in a face image based on discriminant features for face recognition. By analyzing the relationship between the pixels in face images and features extracted from them, the pixels that contain a large amount of discriminative information are selected, while the pixels with less discriminative information are discarded. The experimental results obtained with various face databases show that the proposed pixel selection method results in improved recognition performance, especially in the presence of illumination or facial expression variations. Additionally, the proposed method greatly reduces the memory size and computational load in the face recognition process.

► The proposed method selects important pixels based on a discriminant analysis. ► The proposed method can be simply applied to any type of feature extraction method. ► It can enhance the effects of the selected feature extraction method. ► The proposed method results in improved recognition performance, especially in the presence of various types of variations.