Norihiko MATSUURA, Shohei SUGAWARA and Shigeki MASAKI
NTT Cyber Space Laboratories
1-1, Hikari-no-oka, Yokosuka, 239-0847 JAPAN
TEL: +81-468-59-2822 FAX: +81-468-59-2829
This paper describes the extension of "InterSpace" in order to realize more realistic space for natural way of communication among users. InterSpace is a platform architecture for constructing networked virtual environments using 3D computer graphics ( Sugawara et al. 1994). One of the most important features of InterSpace is that InterSpace integrates the real-time facial video and voice communication capabilities. However, the current InterSpace exhibits shortcomings that limit the potential in computer-mediated communication environment. Especially, we identify the following two areas for necessary improvements:
Avatar Appearances: This involves rich expressions of avatars including human-like avatar animations and so on.
Rich Visual and Audio Communication: This involves high quality real-time facial video and audio transmission, more effective facial video mapping on avatars, rich voice communication, speaker detection method and so on.
Regarding the avatar appearances, human-like representations of avatars is discussed in several approaches ( Benford et al. 1995) ( Benford et al. 1996) ( Pandzic et al. 1995). It is no doubt that the sophisticated human-like avatar representations and body animations are extremely important for natural interaction and communication among participants in networked virtual environments. In most approaches that realizes these demands, facial animations based on body articulation and skin deformations are adopted. In contrast, our system adopted the real-time facial image mapping on faces for interactions,