Key Points
We investigate how an image-based avatar's actions and the rotation of its video affect people's judgments about whom the avatar is addressing in a group. Subjective assessments compare the impressions of the intended recipient and other people.
Abstract
When an on-screen guide addresses a group, people need to know whom it is speaking to. We studied whether movement helps an image-based avatar—a guide made from video of a real person—communicate who should receive its guidance. Earlier work mainly assumed one-to-one interaction.
We tested the avatar's own actions and rotation of the whole avatar video relative to the screen. Participants compared versions with and without these movements. We assessed two different impressions: whether the person requesting guidance felt that the avatar was addressing them, and whether other people understood that the guidance was not for them.
The movements helped the intended recipient feel that they were being addressed. However, they did not help other people recognize that they were not the intended recipient. These results show why an avatar's effect on the person it addresses and its effect on the surrounding people need to be considered separately.
Publications
- Tsubasa Miyauchi, Masashi Nishiyama, Yoshio Iwai,
Directing a Target Person among Multiple Users using the Motion Effects of an Image-based Avatar, Proceedings of 21st International Conference on Human-Computer Interaction (HCII), vol. 3, pp. 341 - 352, July 2019. - Tsubasa Miyauchi, Wataru Ganaha, Masashi Nishiyama, Yoshio Iwai,
Investigation of Motion Video Enhancement for Image-based Avatars on Small Displays, Proceedings of 23rd International Conference on Human-Computer Interaction (HCII), LNCS 12763, pp. 44 - 55, July 2021.
Publication (Japanese)
- 宮内 翼, 西山 正志, 岩井 儀雄,
実写アバタ映像における動き表現を用いた対象者の指定,
ヒューマンインタフェース学会論文誌, Vol. 22, No. 2, pp. 77 - 88, May 2020.
Short movie
