ABSTRACT
It was predominantly the geometric mesh representation or image slices obtained from medical imaging modalities on which information-theoretic measures have found their first applicability in the domain of computer graphics and visualization. Soon after researchers investigated the usefulness of information-theoretic measures in the context of volume visualization. The basic data structure for volume visualization is the data volume, where at various positions in the 3D space, particular characteristics are encoded. The most elementary volumetric representation is an extension of the notion of a grayscale image that is defined on a two-dimensional grid, into 3D, forming a three-dimensional grid. On each grid location, a volume element (voxel) is storing information about the volumetric characteristics of the imaged or simulated phenomenon. Besides this so-called regular volumetric dataset, more complex variants of the volumetric representation may vary in data values over time, or can have multiple values stored per one volume element, or can be organized in an unstructured form instead of a regular-grid alignment.
