Facial Expression Recognition Algorithm

In recent years, the reasons for renewed interest in facial expression recognition are multiple, but mainly due to people face mare interest about human computer interaction (HCI).Facial expression recognition is to analyze and detect the special expression state from given expression images or video frames and then to ascertain the subject’s specific inborn emotion, achieving smarter and more natural interaction between human beings and computers. The study of facial expression recognition has, found important applied values.In this work, we first discuss the background and then analyze the main existing expression recognition algorithms. Then we focus on the face detection and feature extraction. In the face detection part, we study Gray-level integral projection method which used to locate the eye and mouth feature area and we analyze that this method could easily lead to inaccurate positioning which caused by the background impacted. Under the YCrCb color space using skin color extraction, can effectively filter out the background. We do second gray integral projection on the facial image then the feature area can be accurately located.In feature extraction, a new image fusion based on Gabor transform feature extraction method is presented. This method first will send the split the region of interest image into Gabor filters, through5scales and8directions of the Gabor filter, we will get40feature images. These images are fully containing the required identifying features, but some features are redundant features needs to be removed. To reduce the dimension of features, and make full use of all directions and scales characteristics of information of images, image fusion is introduced. Here we used Daubechie wavelet decomposition of image features, on this basis, and took the image fusion of large amplitude and then reconstructed. Not only reduce the dimension of the features, but also effectively extract the face features.Experimental results show that this method can improve the expression recognition rate.