Information Technology Reference
In-Depth Information
Capture and Express Behavior Environment
(CEBE) for Realizing Enculturating
Human-Agent Interaction
Yoshimasa Ohmoto, Akihiro Takahashi, Hiroki Ohashi, and Toyoaki Nishida
Graduate school of Informatics, Kyoto University
Yoshidahonmachi, Sakyo-ku, Kyoto-shi, Kyoto-hu, Japan
{ohmoto@,a.takahashi@ii.ist,ohashi@ii.ist,nishida@}i.kyoto-u.ac.jp
Abstract.
We are studying how Embodied Conversational Agents
(ECAs) express communication behavior with cultural background. The
objective of this study is the proposition of the modified Capture and
Express Behavior Environment (CEBE) in which a person can interact
with ECAs controlled by the captured behavior of another person with
cultural background. In this paper, we discuss modifications and con-
cepts of CEBE to apply CEBE for investigations to realize an ECA with
cultural background. The prototype system could capture basic human
behavior, such as head direction, posture of the upper body, and 3D an-
gles of arms, when each part of the body, such as head, hands, arms and
trunk. In addition, the system could control a robot or a virtual agent
based on the detected data. We have to develop some implementations
to interact with people with cultural background.
Keywords:
human-agent interaction, measuring system, WOZ system,
Robot operation interface.
1
Introduction
In recent years, there has been a growing exchange in goods, information and peo-
ple in the world amidst the progression of internationalization. In cross-cultural
exchange, people can relativize their own values and obtain new values from
different cultures. Although there are some cross-cultural differences in habits
and rules, the experience of recognizing these differences gradually is important
in acquiring communication skill. The reason why people can get though them
is that they have great abilities to learn, adapt and communicate with other
people.
On the other hand, Embodied Conversational Agents (ECAs) which are ex-
pected to provide an effective interface for many users, including computer
novices in diverse fields ([1], [9], [12]), however, have such ability in the distant
future. It is important to interact with people in the fields. When ECAs inter-
act with people, the cross-cultural features influence human-agent interactions.
ECAs can have databases of cross-cultural rules and habits which are known
well in advance. There are however body motions, behavior, and mental stances
