A novel approach for facial expression recognition

Abstract

In this paper, we propose a new facial expression recognition approach based on the use of Gabor filter features. These features are extracted only from regions of interest (ROI). These regions are the components of the face most involved in the expression of emotions. Then we select the best features with PCA. Finally, the SVM classifier is used for the classification of basic facial expressions. The experiment results show the superiority of the method and researched recognition rates of 95.11% and 92.19% for JAFFE and CK+ datasets respectively.