Spoken Emotion Recognition Using Radial Basis Function Neural Network

Recognizing human emotion from speech signals, i.e., spoken emotion recognition, is a new and interesting subject in artificial intelligence field. In this paper we present a new method of spoken emotion recognition based on radial basis function neutral networks (RBFNN). The acoustic features related to human emotion expression are extracted from speech signals and then fed into RBFNN for emotion classification. The performance of RBFNN on spoken emotion recognition task is compared with several existing methods including linear discriminant classifiers (LDC), K-nearest-neighbor (KNN), and C4.5 decision tree. The experimental results on emotional Chinese speech corpus demonstrate the promising performance of RBFNN.