Smile recognition system based on lip corners identification

Automatic smile recognition plays an important part in several intelligent image processing systems. This paper presents an automatic smile detection system based on lip corners identification. Two different corner detection algorithms are used in this paper to identify lip corners, i.e., the Harris corner detection and the FAST corner detection. The proposed system is tested using our VISiO smiling face database. Our results show that the Harris corner detector yields the best result with 77.5% accuracy while the FAST corner detector gives 72.5% accuracy. However, these results depend on the method of of determining p_avg (the average lip corner position during the training phase).