Design an Algorithm for Facial Expression Detection

Emotion is a complex psychological phenomenon and it can be expressed through facial expression. Detecting of this facial expression is very important to interact with people efficiently. Facial expression detection system is a human-computer interaction system to understand human’s expression and emotion. In this paper faces and facial parts are detected, cropped, feature extracted and finally trained individually and as a group. For face detection viola-jones algorithms is used and to detect certain part of the face new method is introduced. For feature extraction local binary patterns, histogram of gradient and Gabor filters are used. Finally classification is done with linear support vector machines. The implementation is done on Japanese Female Facial Expression (JAFFE) Database and Cohn-Kanade (CK+) database and the recognition accuracy is compared between the different feature extraction techniques.