A multi-user framework supporting video-based avatars

In this paper we present our ongoing work in setting up a multi-user framework that supports video texturing on avatars, creating added value for both gaming and collaborative work applications. Limiting the data propagation and bandwidth usage is a key goal when targeting a scalable application that will be deployed on a general-purpose network such as the Internet. We therefore present a number of techniques that can be used in these circumstances, including increased client responsibilities and multicast communications. By making full use of the increased client processing power and bandwidth, we can relieve the server of many of its traditional roles. At the same time, we can target a range of network-enabled devices by handing over control of quality parameters to the individual clients. Through the combination of real-time video distribution and animated avatars, we aim to achieve an increased level of immersion in the virtual world.