Performance evaluation of Local Binary Patterns and it´s derivatives for face recognition

Face recognition is a challenging problem in computer vision and human computer interaction. Texture is the surface property which is used to identify and recognize objects in an image. Texture based facial recognition is a fast growing research area in recent years. The LBP method is based on characterizing the local image texture by local texture patterns. In this paper texture based face recognition is investigated by Local Binary Pattern (LBP) and its derivatives Dominant Local Binary Pattern (DLBP), Local Derivative Pattern (LDP) and Advanced Local Binary Pattern (ALBP). Facial features are extracted and compared using K nearest neighbor classification algorithm. G-statistics distance measure is used for classification. Experiments were carried out on JAFFE female and Yale face databases. The results show that LDP consistently perform much better than the remaining methods.