Semantic Visualization Mapping for Volume Illustration

Scientific visualization is the discipline of automatically rendering images from scientific data. Adequate visual abstractions are important to show relevant information in the data. Visual abstractions are a trade-off between showing detailed information and preventing visual overload. To use visual abstractions for the depiction of data, a mapping from data attributes to visual abstractions is needed. This mapping is called the visualization mapping. This thesis reviews the history of visual abstractions and visualizationmapping in the context of scientific visualization. Later a novel visual abstraction method called caricaturistic visualization is presented. The concept of exaggeration is the visual abstraction used for caricaturistic visualization. Principles from traditional caricatures are used to accentuate salient details of data while sparsely sketching the context. The visual abstractions described in this thesis are inspired by visual art and mostly by traditional illustration techniques. To make effective use of the recently developed visualizationmethods, that imitate illustration techniques, an expressive visualization mapping approach is required. In this thesis a visualization mapping method is investigated that makes explicit use of semantics to describe mappings from data attributes to visual abstractions. The semantic visualization mapping explicitly uses domain semantics and visual abstraction semantics to specify visualization rules. Illustrative visualization results are shown that are achieved with the semantic visualization mapping. The behavior of the automatically rendered interactive illustrations is specified using interaction-dependent visualization rules. Interactions like the change of the viewpoint, or the manipulation of a slicing plane are state of the art in volume visualization. In this thesis a method for more elaborate interaction techniques is presented. The behavior of the illustrations is specified with interaction-dependent rules that are integrated in the semantic visualization mapping approach.