Analysis and investigation of emotion identification in biased emotional talking environments

Abstract

Speakers usually use certain words more frequently in expressing their emotions since they have learned the connection between certain words and their corresponding emotions. The work of this research is devoted to the analysis and investigation of emotion identification in two separate and different talking environments based on classifiers called suprasegmental hidden Markov models. The first talking environment is unbiased towards any emotional state, whereas the second talking environment is biased towards different emotional states. Each emotional talking environment is composed of six emotions. The results of this work show that emotion identification performance in the second talking environment outperforms that in the first talking environment. Based on subjective assessment by human judges, emotion identification performance in the biased talking environment leads that in the unbiased talking environment.