Face Recognition Based on the Feature Fusion of 2DLDA and LBP

To study the robustness of face recognition algorithms on conditions of complex illumination, facial expression and posture, three subset databases (Illumination, Expression and Posture subsets) are constructed by selecting images from several existing face databases. Advantages and disadvantages of seven typical algorithms on extracting global and local features are discussed respectively through the experiments on ORL and the three databases mentioned above. To improve the recognition rate, an algorithm of face recognition based on the feature fusion of Two-Dimensional Linear Discriminant Analysis (2DLDA) and Local Binary Pattern (LBP) is proposed in this paper. The experimental results verify both the complementarities of the two kinds of feature and the effectiveness of the proposed feature fusion algorithm.