Emotional State Analysis Through InfraRed Thermal Imaging

Abstract

Human-machine interaction has growing interest in studies about recognition of human emotions since these are relevant to the establishment of social relationships. Some markers of human emotions are physiological signals, such as heartbeat, brain signals, sweating and skin temperature. Many of them are captured through sensors with contact with the body, however, contact-free (unobtrusive) sensors are currently studied targeting those people who have sensitivity to touch, such as the case of people with Autism Spectrum Disorder (ASD). InfraRed Thermal Imaging (IRTI) is an unobtrusive technique for recording thermal variations on the skin. The goal of this work is to investigate the thermal variation from two emotions (happiness and sadness) on the face of five typically developing children through IRTI. The facial regions of interest (ROI) are forehead, cheeks, tip of nose and periorbital. This investigation is based on the brightness variation between the exposition to two emotional stimuli (evoked by videos) and the absence of stimuli (baseline). Values from Student’s t-test indicated significant differences between both emotional states (positive/happiness, and negative/sadness) and the baseline.