Impact of emotion type on emotion recognition through vocal channel

The topic of emotion in computing is enjoying recent and growing attention. Until recently, the role of speech to understand emotion has been frequently ignored. In this paper, we have developed an emotion recognition system based on vocal channel highlighting the impact of the link between the features used for characterizing emotion state and the nature of emotion. For the purpose of this work, we have examined several features extraction methods (MFCC, LFCC and Energy) applied with Hidden Markov Model (HMM) as classification system. The performance of the proposed approach is evaluated on real condition speech signal (IEMOCAP database). It is shown that the recognition of emotion varies significantly depending on the nature of emotion. Furthermore, we demonstrate that each emotion type is better characterized by a different type through the battery of speech feature used.