(英) |
This paper presents a method to create an individual model to describe relations between facial expressions and stress patterns using Bayesisan networks. For creating this model, we extract relations between three facial expressions (happiness, anger, and sadness) for the basic six facial expressions defined by Ekman and psychological stress factors (dysphoria and anxiety, displeasure and anger, and lassitude). In the evaluation experiment, we constructed relational models of individual facial expressions from the dataset of facial images and stress vales measured using Stress Response Scale-18 (SRS-18) obtained from 10 subjects during 7-20 weeks at one-week intervals. Moreover, we demonstrate psychological stress factors on facial expressions with using to estimate probabilities using arousal levels of facial expressions. |