Abstract:Visualization has become a primary way of interpreting vast amounts of data generated by acquisition and computational simulation techniques. Large complex datasets with high dynamic range and numerical precision pose challenges for successful visualization. It is also required that users take an active role in the visualization process for better data interpretation. In this work, we have developed strategies and methods to address both effective and efficient visualization. They fall into three components: high fidelity visualization, illustrative visualization and explorative visualization. First, to achieve high fidelity visualization, we develop a framework for high dynamic range volume visualization on PCs with commodity graphics hardware to faithfully render high resolution volume data sets with wide numerical ranges. We further introduce a method inspired by interference phenomena to visualize surface deformation with large dynamic range. Secondly, to achieve illustrative visualization, we present several nonphotorealistic rendering (NPR) techniques for surface illustration including Procedural Image Processing (PIP), point-based silhouette rendering, and Shading by Random Number (SRD). We also develop a novel framework for stippling and silhouette rendering operating in the geometry image domain, which provides combined benefits of the existing image space approaches. These surface illustration techniques can be combined with direct volume rendering to effectively illustrate surfaces in volume. Lastly, to achieve explorative visualization, we have developed sketch-based segmentation of volume date sets. Our experiments show that our method extracts volumetric structures precisely and efficiently while requiring little skill or effort from the user. Finally, based on the work we have successfully developed, directions for future work are presented in the last part of this thesis.

High fidelity, illustrative and explorative visualization