The Research on Cross-Language Emotion Recognition Algorithm for Hearing Aid

Xia Shulan, Wang Jilin

COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(12C) 379-383

College of Electrical Engineering , Yancheng Institute of Technology , Yancheng 224051 , Jiangsu, PR China

To improve the emotional perception of hearing-impaired people, the speech emotion recognition cross language algorithm is proposed. The mutual features of speech emotions in the Chinese and German corpora are analyzed and a speech emotion recognition algorithm using the Restricted Boltzmann Machine based on Parallel Tempering is proposed. First, the algorithm calculates 375 acoustic features. Then, by using the method of Fisher Discrimination Ratio combined with weighted feature fusion, 20 features are selected to be used in the speech emotion recognition. Afterwards, it sets up the Restricted Boltzmann Machine to recognize 5 types of speech emotions (anger, fear, happiness, neutral, sadness). The experiments show that this algorithm can effectively recognize the cross-language speech emotions except the neutral emotion.