Face Recognition Based on Deep Belief Network Combined with Center-Symmetric Local Binary Pattern

Human face recognition performances usually drops heavily due to pose variation and other factors. The representative deep learning method Deep Belief Network (DBN) has been proven to be an effective method to extract information-rich features of face image for recognition. However the DBN usually ignore the local features of image which are proven to be important for face recognition. Hence, this paper proposed a novel approach combined with local feature Center-Symmetric Local Binary Pattern (CS-LBP) and DBN. CS-LBP is applied to extract local texture features of face image. Then the extracted features are used as the input of Deep Belief Network instead of face image. The network structure and parameters are trained to obtain the final network model for recognition. A large amount of experiments are conducted on the ORL face database, and the experimental results show that compared with LBP, LBP combined with DBN and DBN, the proposed method has a significant improvement on recognition rates and can be a feasible way to combat with pose variation.