Reconhecimento de emoções: um estudo de caso utilizando biossensores

Abstract

The goal of this research is to develop a user independent emotion classifier to infer the emotions of happiness, anger, sadness and pleasure, by using biosensors. In order to achieve this goal, the electrodermal activity, peripheral temperature and blood volume pulse were collected from 15 subjects ( 10 males and 7 females). These signals were collected by a hardware built for this research. A software was developed to exhibit the emotional stimuli, manage the exposition time, transmit and store the data from the hardware. Also, the experiment allowed to apply tests to label data and capture pictures to be used to create labels when submitted to a facial emotion recognition system. As a preliminary result, the classifier achieved similar results compared to those available in literature, when concerning user independence data (sic).