Combining LBP and Adaboost for facial expression recognition

Abstract

A novel approach to facial expression recognition based on the combination of local binary pattern (LBP) and Adaboost is proposed. Firstly, facial expression images are processed with LBP operator, which can eliminate the effect of environment lighting in a certain extent and has the powerful capability of texture feature description. And then facial expression features are presented with LBP histograms of expression image which is divided into several blocks. The features with powerful discriminability are selected by a modified Adaboost so as to predigest the design of classifier and shorten the cost time. Finally, the support vector machine (SVM) classifier is used for expression classification. The algorithm is implemented with Matlab and experimented on Japanese female facial expression database (JAFFE database). A facial expression recognition rate of 65.71% for person-independent is obtained and shows the effectiveness of the proposed algorithm.