Survey on Intelligent Speech Emotion Recognition

Abstract

Speech emotion recognition is the most representative branch of affective computing, which is of great significance for achieving harmonious human-computer interaction. This paper systematically surveys intelligent speech emotion recognition algorithms, including speech emotion datasets, classical feature extraction and dimensionality reduction, and various leading emotion classification models. By enumerating the excellent works of speech emotion recognition in the past 20 years, the pros and cons of these works are compared and analyzed. Finally, we summarized the main problems of speech emotion recognition and the future directions, in order to promote human-computer interaction to a more humane stage.