Control structure and multi-resolution techniques for virtual human representation

Abstract

A virtual human is a typical instance of articulated physical objects: it does not have only one shape but many, corresponding to all the possible postures that the underlying articulated skeleton can reach. For realistic rendering results, a high-quality texture is usually associated to the shape and skeleton structure. Controlling and animating a virtual human model requires simultaneously many graphics and computational resources. The first part of this chapter discusses the control articulated skeleton structure and different approaches to build skeletons [10] and bind it to the geometry. The second part addresses the production of LoDs for virtual humans, both for the 3D shape (geometry) and the articulated skeleton (motion and animation).