Development of emotion recognition algorithm using tactual communication tool

Some tactual communication tools which conveys emotions during face-to-face oral conversation have been developing. In the present study, to diagnose the mental condition of the institutionalized elderly person using the tactual communication tool, an algorithm of emotion recognition was developed. The emotion recognition algorithm recognized six emotions (surprise, joy, anger, sadness, disgust and fear) as three emotion groups (surprise; joy and anger; sadness, disgust and fear). Only four parameters (the number of grips, maximum voltage, release time of one-time grip and mean of the release and grip time of multiple grips) were selected for emotion recognition. Three cases of one-time grip, 2–8 times grips and over 9 times grips were considered. Even though the algorithm was very simple, the overall accuracy of the recognition of the emotion group was 73.5%. Therefore, it was considered that the derived emotion recognition algorithm was effective.