Emotion recognition — An approach to identify the terrorist

Abstract

The emotional influence on human behavior can be identified by speech. Recognition of emotion plays a vital role in many fields such as automatic emotion recognition etc. In this paper, we distinguish a normal person from the terrorist/victim by identifying their emotional state from speech. Emotional states dealt with in this paper are neutral, sad, anger, fear, etc. Two different algorithm of pitch is used to extract the pitch here. Moreover, support vector machine is used to classify the emotional state. The accuracy level of the classifier differentiates the emotional state of the normal person from the terrorist/victim. For the classification of all emotions, the average accuracy of both male and female is 80%.