Facial Expression Recognition Based on Local Binary Patterns and Least Squares Support Vector Machines

Abstract

In this paper a new facial expression recognition method based on local binary patterns (LBP) and least squares support vector machines (LS-SVM) is proposed. LBP is adopted as facial representations for facial expression recognition since LBP tolerates against illumination changes and operates with its computational simplicity. After extracting LBP features, LS-SVM with radial basis function (RBF) kernel is employed for facial expression classification. The experimental results on the popular JAFFE facial expression database demonstrate that the recognition accuracy based on LBP and LS-SVM comes up to 78.57%.