High-Level User Interfaces for Transfer Function Design based on Semantic Models

From the technological point of view, a variety of new and useful features has been added to volume rendering algorithms in recent years, including multi-dimensional transfer functions and sophisticated shading models. However, advanced volume rendering techniques are not widely used in practice. This is mainly due not to technological reasons, but to difficulties in managing the complexity of visual parameter assignment which is done by means of transfer functions. Users report on problems concerning the process of specifying optical properties for datasets. On the one hand, automatic approaches of designing a transfer function are often not adaptable and flexible enough in order to realize a desired visualization. On the other hand, manual transfer function assignment is a challenging task even for expert-users, since appropriate user interfaces, especially for multi-dimensional transfer functions, are difficult to operate. This thesis proposes an approach of transfer function design which deals with semantics. For this purpose, an additional abstraction layer for parametric models of transfer functions is introduced which facilitates the specification of optical properties. The method of calculation is based on principal component analysis and adapts concepts from the field of computer animation. Additionally, a framework is presented which allows visualization experts to design high-level transfer function models which can be used intuitively by non-expert users. As a result, user interfaces are obtained which provide semantic information for specific application areas. Within this thesis, two semantic transfer function models for medical visualization are developed, namely Magnetic Resonance Imaging and Computed Tomography Angiography. Besides the practical work, the following aspects are discussed in this thesis: an overview of the basic techniques which this work implies, the development of a theoretical semantic transfer function model, the practical realization and implementation of the transfer function model into a framework, and the visualization of the results based on semantic models.