C209 鼻部熱画像処理に基づいた感情の可視化

抄録

We studied visualization of emotions using facial thermal images. We constructed pleasure-feeling (P-U) estimation model and arousal-feeling (A-S) estimation mode 1 based on nasal temperature analysis. Then, we attempted emotion-estimation on watching video programs. Four video programs were prepared for 4 subjects. P-U and A-S estimate were analyzed from thermal images, and the vectors were plot ted. Then we observed that FEAR was distributed locally on the two quadrant of the plane. And in the other three cases, LAUGHTER, ANGER and SADNESS, th ere were controlled by P-U or A-S. As the result, P-U and A-S were important parameters for the emotion-estimation.