Improving the Face Recognition Accuracy under Varying Illumination Conditions for Local Binary Patterns and Local Ternary Patterns Based on Weber-Face and Singular Value Decomposition

This paper addresses a new approach based on the Weber-face and singular value decomposition (SVD) methods to improve the recognition accuracy for a face recognition system using local binary patterns and local ternary patterns in an illumination variation environment. The face images are the first extracted illumination-invariant components by the Weber-face method. Secondly, SVD is applied to the encoded images. Next, the training encoded images are extracted features based on local binary patterns or local ternary patterns. Finally, in the classification phase, the singular value matrix of a test image is combined with those of the training images to adjust the illumination of the test image before the features are extracted and classified. The recognition is performed using a nearest neighbor classifier with Chi-square as a dissimilarity measure. Experimental results on the extended Yale B database demonstrated the efficiency of our proposed method. Thus, the proposed approach is expected to contribute to the face recognition problem under varying illumination conditions.