Real-time facial expressions identification using supervised descent method

Human's facial expressions are related to the position of the facial features. In this paper, we use Supervised Descent Method to find the position of the features which converges several times by a template and can let the feature points to be on the right place. After getting all the feature points, we find out the relationship between these points and the facial expressions. At the meanwhile, we also need the neutral face of the users to be used as the basis for identification.