A New Facial Expression Recognition Method Based on Local Gabor Filter Bank and PCA plus LDA

ses a facial expression recognition system based on Gabor feature using a novel r bank. Traditionally, a global Gabor filter bank with 5 frequencies and 8 ten used to extract the Gabor feature. A lot of time will be involved to extract mensions of such Gabor feature vector are prohibitively high. A novel local Gabor art of frequency and orientation parameters is proposed. In order to evaluate the he local Gabor filter bank, we first employed a two-stage feature compression LDA to select and compress the Gabor feature, then adopted minimum distance nize facial expression. Experimental results show that the method is effective for eduction and good recognition performance in comparison with traditional entire . The best average recognition rate achieves 97.33% for JAFFE facial expression abor filter bank, feature extraction, PCA, LDA, facial expression recognition. deliver rich information about human emotion and play an essential role in human In order to facilitate a more intelligent and natural human machine interface of new cts, automatic facial expression recognition [1][18][20] had been studied world en years, which has become a very active research area in computer vision and n. There are many approaches have been proposed for facial expression analysis ages and image sequences [12][18] in the literature. we focus on the recognition of facial expression from single digital images with feature extraction. A number of approaches have been developed for extracting by: Motorola Labs Research Foundation (No.303D804372), NSFC (No.60275005), GDNSF 105938).