Robust Face Recognition Approaches Using PCA, ICA, LDA Based on DWT, and SVM Algorithms

Face Recognition represents one of the attracting research areas. It has drawn the attention of many researchers due to its varying applications such as database identification, identity authentication, surveillance and security access control human computer interaction.... In this order, different face recognition algorithms are proposed. These methods are based on the face representations. In order to use the most important information in the face representation, we propose to exploit the concept of 2D-DWT for the image compression as a preprocessing for feature extraction. In fact, the DWT is performed at different scales and orientations so it is sensitive to strong lighting conditions and facial details. Based on the DWT advantages, the LL sub-band of the processed image is used as input image for feature extraction process based on ICA, PCA, LDA and SVM algorithms. The importance of the proposed approaches is evaluated by experimental and comparative studies tested on AT&T database.