Dynamic recognition of basic facial expressions by discrete-time recurrent neural network

In order to develop an "active human interface (AHI)" that realizes heart-to-heart communication between machine and human being, the authors have been investigating the recognition of human emotions from facial expressions. This paper deals with the recognition method of dynamic facial expressions. The authors construct the recurrent neural network equipped with a special function to change the reference viewpoint in calculating facial information for machine recognition of dynamic facial expressions, and then, the authors find that the recurrent neural network can perform almost similar recognition to that of human beings.