Scientists use visualizations to better understand their research data and communicate their findings with others. Making visualizations can be a tedious task for the scientists if the task requires a complex process demanding frequent user intervention. An example is volume data visualization, which requires the user to specify a set of filtering, classification, and rendering parameters and functions. Existing volume visualization systems come with rather sophisticated, task-specific user interfaces for each stage of the volume visualization process. Learning to operate these interfaces is not intuitive, and can be unpleasant. This dissertation makes a contribution to the development of intelligent visual interfaces for volume visualization. The idea is to employ artificial intelligent systems to carry out some of the visualization tasks for the user such that the visual interface is greatly simplified making the overall system more intuitive and easier to use. Both supervised and unsupervised intelligent systems are studied for typical volume segmentation and visualization tasks required by a variety of scientific and medical applications. My research results can potentially become a basis of the design and development of next-generation user interfaces for data visualization.
展开▼
