Feast: face and emotion analysis system for smart tablets

Face and emotion recognition is still an open and very challenging problem. This paper presents a system FEAST which is an intelligent control system of Smart Tablets. It involves manipulating user sessions to adapt the working environment to his emotional state. First, a face detection followed by face and emotion recognition is performed, then a profile change is made basing on the obtained results. The face detection is based on skin color and geometric moments and face recognition is done by merging two features spaces, namely, Zernike moments and EAR-LBP. A feature selection technique reducing the parameter space size is applied. The same parameters are used for the emotion recognition.