Facial Expression Recognition Based on Supervised LLE Analysis of Optical Flow and Ratio Image

Abstract

In this paper, we propose a new facial expression recognition algorithm based on supervised locally linear embedding (SLLE) analysis on the optical flow and ratio image. In this algorithm, we first extract the face region from the face image to remove factors due to global head motion. Secondly, we compute the optical flow and ratio image between the neutral face and expression images and then apply the SLLE to extract the low-dimensional discriminating features from the expression motion and brightness variation. Finally, we compute the distance between the low-dimensional feature vectors to recognize the facial expression. Experimental results on the JAFFE face database and Yale database show that the proposed algorithm outperforms previous methods for facial expression recognition.