Expression Recognition Using RP and DPL

Abstract

Expression representation and classifier design are two vital aspects of automatic facial expression recognition. Using different Random Projection (RP) for each class features and Dictionary Pairs Learning (DPL) classification algorithm, the proposed approach of face expression recognition can classify the prototypic emotional facial expressions with improved computation burden and recognition performance. With the computation complex can be greatly decreased, the experimental results show that the provided method achieves the almost perfect hit hate both for MUG and MultiPIE databases.