Information Technology Reference
In-Depth Information
Fig. 6. The examples of captured data of faces and hands by an IR-range sensor
Facial expressions. There are many software which convert facial features of a
person into facial features of an avatar. We can easily invoke the external program
as a plug-in from the virtual agent which we use because the virtual agent are
worked on GECA Framework. When we use the software which can export facial
features in real time, the virtual agent can express the facial expressions of the
user in CEBE.
Hand gestures. We did not implement the method to capture hand gestures
because we preferentially implemented the methods to capture human behavior
to apply CEBE for using robots. It is dicult to capture hand gestures by using
a motion capture system needed makers. We expect that we can capture hand
gestures because we can obtain 3D positions of the surface by using IR-range
sensors.
Ghobadi et al. [3] proposed a system to control hand based robot using IR-
range sensors. Figure 6 also shows that IR-range sensors can capture the features
of a hand. In addition, we can easily trace the hand positions because we estimate
3D-angles of arms. We thus expect that we can capture more detailed hand
gestures by using both IR-range sensors and pan-tilt cameras.
5.2 Limitations
We will have to develop some implementations so that an agent reflects human
behavior and can interact with people with cultural background.
It is most important for natural interaction to compensate for the delay time
of agent actions. This delay was pronounced when a robot was controlled. To
resolve this issue, we will have to implement a predictive control mechanism. On
the other hand, we need to investigate the influence of the delay time of agent
actions in encultured agent-human interaction.
Search WWH ::




Custom Search