Improved facial expression recognition using graph signal processing

Abstract

A novel method of facial expression recognition using graph signal processing (GSP) is proposed. Here the dimensionality of the feature vector of the facial expression based on the histogram of oriented gradients and discrete wavelet transform is reduced using the GSP and then applied to a classifier. The proposed scheme is compared with the existing methods on the JAFFE dataset in terms of the accuracy and the length of the feature vector. It is seen from the simulation results that the proposed scheme provides better accuracy than the existing methods.