Application of Completed Local Binary Pattern for Facial Expression Recognition on Gabor Filtered Facial Images

Automatic facial expression analysis is vital in identifying the emotional state of a human being. Hence it is an essential tool for human-computer interaction. Success of a facial expression recognition system largely depends on selecting suitable and effective facial features. This paper proposes a novel approach for recognizing facial expression where facial feature is represented by completed local binary pattern (CLBP) applied on Gabor filtered facial image. Expression images are classified into prototype expression via support vector machine (SVM) with different kernels. Experiments performed on Cohn-Kanade facial expression database obtained a recognition rate of above 97% for seven-class expression set indicating superiority of the proposed method compared to other methods.