ABSTRACT
Face recognition is still a challenging issue especially across the non-uniform motion blur, illumination and pose. There are several approaches are available for face recognition in that deep learning approaches are most popular. The Deep Belief Network (DBN) is an efficient deep learning method to extract the features from the face image. However, in face recognition, local features play a prominent role to recognize the faces but DBN usually ignores the local features. To overcome these problems we propose a Local Binary Pattern (LBP-DBN) approach. The LBP is used to extract the local features. The features extracted from LBP are feed to DBN and then obtain a final network model for recognition. The proposed method is carried out on a standard face recognition datasets like AT&T. From the experimental results, we can say that the proposed method gives better performance compared to other state-of-art techniques.
