Classifier fusion for speech emotion recognition

Abstract

According to multidimensional emotion space model, an improved queuing voting algorithm was proposed to implement the fusion among multiple emotion classifiers for a good emotion recognition result. Firstly, three kinds of classifier were designed based on hidden Markov model (HMM) and artificial neural network (ANN). Then, the improved queuing voting algorithm was used to fuse them. Experimental study had been carried out using Beihang University mandarin emotion speech database and Berlin database of emotional speech respectively. The results proved that the improved queuing voting algorithm can attain better fusion effect than conventional fusion algorithm and excel any single classifier evidently.