ILLUMINATION INVARIANT FACE RECOGNITION

Abstract

Face recognition is one of the most important tasks in computer vision and Biometrics. LBP being a powerful feature for texture descriptor useful for identifying objects or regions of interest in an image. The texture based face recognition is widely used in various applications. The local binary pattern (LBP) method is one of the most successful for face recognition. It is basically based on characterizing the local image texture by local texture patterns. The performance evaluation of Local Binary Pattern (LBP) on simple faces, faces with occlusion and faces with illumination variation. The facial image is divided into local regions and texture descriptors are extracted from each region independently. These facial descriptors are then concatenated to form a global description of the face. The LBP labels for the histogram contain information about the patterns on a pixel-level. These labels are combined over a small region to produce information on a regional level The regional histograms are concatenated to build a global description of the face. Facial features are extracted and compared using K nearest neighbour classification algorithm. Eucledian distance measure is used for classification. The results shows that LBP consistently performs much better than the remaining other models. In the presented work we have proposed a novel approach for face recognition using LOCAL BINARY PATTERNS. For testing the proposed approach, the ORL (Olivetty Research Laboratories) database is used. It consists of 20 subjects, each having 8 orientations.