Facial Expression Recognition Based on Feature Fusion and Fuzzy Kernel Discriminant Analysis
Xiaoyan Zhou · 2009
In this paper,a facial expression recognition method based on feature fusion and fuzzy kernel discriminant analysis (FKDA)is proposed. This method firstly locates 34 landmark points from each facial image as the Geometric features of the facial image. Then,these landmark points are converted into a labeled graph (LG)vector using the Gabor wavelet transformation method,and the LG vector are used as the Gabor feature vector of the facial image. Both Geometric feature and Gabor feature are further fused using the canonical correlation analysis (CCA)as the final input facial features for recognition. The FKDA method is finally used to further extract the discriminative expression features for classification and the nearest neighbor classifier is used to this goal. Experiments on both Japanese Female Facial Expression (JAFFE)database and the Ekman's 'Pictures of Facial Affect' database demonstrate the better performance of the proposed method.