Development and Analysis of Multimodal Biometrics System for Emotion Detection

Abstract

Various emotions are produced by the newlinehuman during interaction or communication newlineand they vary in meaning, intensity, and newlinecomplexity. Behavioral Emotions can also newlinebe expressed by human body language, newlinevoice tone, electroencephalography, facial newlineexpression, and physiological data facial newlineexpression, etc. and it may be easy to newlineunderstand, but when we are talking about newlinethose humans which not able to express their newlineemotions by behavior, in this situation newlineemotion can be recognized by their brain newlinesignal. newline newlineIn this research, we classified the newlineemotion using a facial expression and newlinesignal which produces in the human brain newlineaccording to their emotional state. We newlineused for facial expression, median newlinefilter, for preprocessing, Gabor filter and newlinePCA is used for feature extraction, and newlineneural network is applied for the newlineclassifier. EEG eight frequency band newlinesignals for emotion detection. STFT is newlineused for feature extraction and newlineclassification SVM is applied. At newlinedecision level sum rule and product rule newline newlinemethod is applied on facial emotion and newlineBrain signal emotions for calculating the newlineresults. newline newlineExpressions were discussed and newlinerecognized different emotions such as newlinehappy, sad, angry, fear, surprised, newlineneutral, and disgust by showing different newlineemotional movie clipping and analyzing newlinethe intensity of the emotion by the facial newlineexpression and brain signal and newlinecomparing it with the statement of the newlineperson. For this research we have newlineprepared the dataset of 72 people and 42 newlinedifferent emotion-based movies clipping newlinefor emotions captured. newline