Research Progress of Speech Emotion Recognition Based on Discrete Emotion Model

Abstract

Speech emotion recognition is one of the important technologies of human-computer interaction, and neural networks have made great contributions in it. In this survey, the commonly used discrete emotion databases and speech emotion feature parameters are firstly introduced. Then the feature extraction methods and emotion recognition models in recent three years’ research in China are summarized, and the former one can be divided into two directions, selecting the feature subset from the existing features and using neural network to extract new features. Finally, the challenges encountered in speech emotion recognition are described, and the future research and development of speech emotion recognition based on discrete emotion models are prospected.