ABSTRACT
IN SCIENTIFIC SCENARIOS, VOLUMETRIC DATA is frequently obtained by 3D measurement, such as computerized tomography (CT) or magnetic resonance imaging (MRI). The volumetric objects used for visual arts and animation usually do not exist in the real world. They cannot be simply obtained by scanning an existing object. This chapter deals with modeling techniques that support the artist in creating volumetric models from scratch. Voxelization techniques allow volumetric representations to be created from polygonal surface descriptions by spatial discretization. Procedural modeling techniques are used to describe internal structures. Similar to photographed texture images for polygonal models, real 3D image data acquired by tomographic measurement can be used to supplement volumetric models.
As an example, Figure 12.1 illustrates the creation of a volumetric model for a three-legged creature. The outer shell of the creature as well as the inner organs, such as bones and muscles, are created using traditional surface modeling techniques. For internal structures, both procedural volumes and measured data has been used. 3D image processing operations, such as blurring and edge enhancement, are used on the voxelized polygonal models. All individual textures are finally combined using 3D compositing techniques. Finally, transfer functions are applied to define the optical properties inside the volume, and the model is rendered using a specified shading model.
