Face recognition based on LBP and CS-LBP technique under different emotions

Abstract

Recognition of a face under different emotions is a challenging subject. Human expression frequently changes with mood. during recognition of face, human may have any expression (neutral, happy, sad, angry, disgust or surprise) and thus, in case, where only neutral face expression has been stored in database, it become difficult to recognize the face of person which is under emotions other than neutral expression. The work done in this paper is twofold. First, local binary pattern (LBP) and center-symmetric local binary pattern (CS-LBP) has been applied to extract the local binary features of the image. Second, Euclidean distance, histogram intersection and chi-square distance is used for recognition of face. The performance is evaluated on the Japanese Female Facial Expression (JAFFE) database and results are compared in terms of recognition rate and time taken for processing. It has been observed that CS-LBP provides better recognition rate rather than LBP in case of different expressions of face.