Emotion Analysis using FST

Abstract

The emotions and autonomous nerve system are physiologically closely related, as are emotions and facial skin temperature because changes in FST are caused by vein constrictions. We studied objective emotion discrimination using facial skin thermography. The emotions are equanimity, fear, angry, joy and sadness. Measurement results showed that FST increased in anger, joy, and sadness but decreased in fear and that nasal FST changes the most, but it is difficult to discriminate between feelings by measurement results. This paper describes feeling by thermogram of the nose based on a neural network. In this experiment, the FST among test subjects made the different of skin temperature's distance. For a solution, the method to make use of input data of Neural Network by formal data. Before of this problem, Neural Network did not converge. After solution, it attained a recognition rate of 71.7%. This paper shows that this system is efficient. It showed that there is some possibility of quantitative evaluation of feelings and of various application of the system in the future.