G-13: Eye-Gaze Control of Virtual Agents Compensating Mona Lisa Effect
| 発表番号 | G-13 |
| 発表タイトル | Eye-Gaze Control of Virtual Agents Compensating Mona Lisa Effect |
| 筆者氏名・所属 |
Hsueh-Han Wu（東京工業大学大学院 情報通信系） Hironori Mitake（東京工業大学大学院 情報通信系） Shoichi Hasegawa（東京工業大学大学院 情報通信系） |
| アブストラクト | The uses of digital signage with a virtual agent have gradually increased nowadays. Like real people do eye-contact to keep others’ attention or to have communication，we also expect the virtual agent to exert the functions of gaze on the viewers. However，the 2D display will lead to the inaccurate visual perception of gaze direction. This phenomenon is referred to as the “Mona Lisa” effect，which the gaze of a portrait appears to follow observers no matter what viewing angle they stand.Our goal is to design a system that the virtual agent makes eye contact with observers in the necessary timing. We consider that there are mainly 2 situations while we look at the virtual agents. One is standing still to look，and another is looking while passing by the display. We conducted the experiments in these 2 situations separately. By controlling iris position and head rotation of the stimuli，we have participants to judge whether they feel the eye contact with the virtual agent. |
| 論文 | PDFファイル |
