Summary
Numerous Non-Player Characters (NPCs) are featured in video games, and some of the humanoid NPCs have the role of shaping the urban landscape by wandering through the streets or resting on benches. However, many of them exhibit only mechanically behaviors, predetermined by creators, resulting in a lack of naturalness in the scenery compared to the real world.
In this study, we employ an A-V model, representing human emotions along two axes: Arousal and Valence, to generate animations reflecting the emotions of NPCs. By setting environmental and surrounding human emotions as parameters, we aim to depict the emotional changes of individual NPCs, and generate natural scenes with the entire crowd.
Members
| Name | Affiliation | Web site |
|---|---|---|
| Ryo Oji | Keio University | |
| Ruizhe Li | Keio University | |
 Masanori Nakayama | Keio University |
Video
Preparing
Publications
Presentations
- Ryo Oji, Li Ruizhe, Masanori Nakayama, Issei Fujishiro: “Generating Naturally Animated Crowd Scenes with Non-Player Characters Based on Arousal-Valence Model―Reproduction of audience interaction－,” in The Technical Report of the Institute of Image Information and Television Engineers, Vol. 45, No. 8, pp. 33―36, March, 2021, Excellent Presentation Award (in Japanese).
- Ryo Oji, Li Ruizhe, Masanori Nakayama, Issei Fujishiro: “Generating Naturally Animated Crowd Scenes with Non-Player Characters Based on Arousal-Valence Model,” in Proceedings of the 83th National Convention of International Processing Society of Japan, Vol. 4, pp. 95―96 (7Y-05), Online, March 18―20, 2021 (in Japanese).
Grants
Grant-in-Aid for Scientific Research (A): 17H00737 (2020―2021)