Facial Expression Recognition Based on PCA-SIFT and Fuzzy-Edge Detection.

The traditional method of facial expression recognition (FER) is difficult to accurately describe a facial expression. The illumination, facial angle, individual differences, and other factors make the low recognition accuracy. Thus, in this paper, we present a method of combining SIFT to filter the feature points and PCA to reduce the dimension of feature images. Meanwhile, fuzzy edge detection is adopted to detect the edge of face. For the experimental result, this method outperforms other state-of-the-art methods