Facial expression recognition in video sequence images by using optical flow

In this paper a new method for facial expression recognition is presented. According to this algorithm, an appropriate mask is designed using Gabor filters, and it is convolved with first frame of video sequence images. Then oval part of face is specified and its main components are characterized. By using Lucas Kanade method for optical flow analysis to determine the motion flow vectors on the regions of main parts during frames largest, motion vectors related to sensitive points of face are extracted and classified to the six basic classes such as: normality, happiness, sadness, anger, disgust, and surprise, facial expression are extracted. This method has high accuracy in comparison with other methods and don´t need for select landmark manually at first.