Analysis of Emotions through the Use of Physiological Signals

Abstract

Emotion is a psychosomatic process that is caused either by a conscious or non-conscious perception of an object or a situation whereas, the result of that process is depicted either by the expression of the face, or physically, or with a combination of the above. An important mechanism playing a significant role in the recognition of emotions is the encephalogram, which enables us to detect significant signals of the brain. The methodology developed aims to the identification of basic emotions by using an electroencephalogram through an experimental process where participants were asked to evaluate the experiment according to the valence, arousal and dominance of each emotion inflicted on them by videos they were subjected to watching. Through the experimental process, a vector of characteristics is analyzed for each of the electrodes of the electroencephalogram and for a series of videos. The training of a model and classification of algorithms were the key to finding the best methods in terms of sensitivity, accuracy, and specification of our data.