Multimodal analysis of human fear

Abstract

Human emotion detection is very much relevant in today’s scenario. Human life become fast due to modernisation. People like to lead sophisticated, peaceful and healthy life. Human emotion plays a vital role in present scenario. Six basic emotions are considered for research purpose. Those are happy, sad, fear, anger, disgust and boredom. In this paper, human fear is analysed based on Electroencephalogram (EEG) signal, physical parameters and facial images. Statistical parameters both from time and frequency domain are used as feature set. Own database is used for the analysis. It is seen from the result that the efficiency is enhanced significantly after multimodal analysis of human fear. The classification results for discrete wavelet transform and logistic regression model are improved by 8.33% and 8.33% respectively.