Visual emotion recognition based on dynamic models

Abstract

This paper introduces the semi-continuous Hidden Markov Model (HMM) and proposes a novel Dynamic Bayesian Network (DBN) model for dynamic visual emotion recognition. The former reduces the training complexity caused by mixture Gaussians by sharing the Condition Probability Densities (CPDs) of Gaussians among the states, and the latter adds a sub-state layer between state and observation layer based on traditional DBN framework and describes the dynamic process of visual emotion in detail. Experiments results show that semi-continuous HMM and three-layer DBN have better performance, and average emotion recognition rate of the semi-continuous HMM is 1.85% and 3.82% higher than those of classical HMM and mixture Gaussian HMM respectively, and average emotion recognition rate of three-layer DBN is 1.93% higher than that of traditional DBN.