Host: The Japanese Society for Artificial Intelligence
Name : The 33rd Annual Conference of the Japanese Society for Artificial Intelligence, 2019
Number : 33
Location : [in Japanese]
Date : June 04, 2019 - June 07, 2019
Expressing emotion is important for communication on the Internet, and some avatars can express their facial expressions. However, it is difficult to create facial animations of an avatar to express complicated emotional states (e.g. happy but embarrassed) for non-expert creators because there is no formulated method proved scientifically. This paper proposes a method to create facial animations to express “suppressed positive emotion” so called “Tsundere” by using combination of action units. In order to apply two types of suppression methods, “neutralized” and “masked,” the action units are chosen by considering “suppressing or suppressed emotion” and “controllable or uncontrollable unit.” 18 facial animations are created and evaluated by Sheffé’s method. As a result, a combination of slight-happy action units and blushing and that of anger-shaped eyebrows and blushing were the most effective to express “suppressed positive emotion.”