Volume Exploration Made Easy Using Feature Maps

We present a framework that enables an intuitive, feature-centric exploration of segmented volumetric datasets. Our system is geared towards users familiar with the basic elements of volume rendering, but who seek to conduct volume exploration in a guided fashion. It provides the infrastructure to organize features and objects extracted from a volume dataset, via segmentation or otherwise, and provides the functionality to view these features with standard volume rendering tools. A novel aspect of our system is that it does not require a separate binary tag volume to indicate the presence of a feature (or object). Instead, we mark a feature by migrating its density range, including its smooth boundary, to a private interval. This avoids the aliasing problems associated with binary tag volumes as well as the extensive run-time costs incurred to resolve these. In addition, since the smooth boundaries of the features are preserved, any volume renderer can be used for data visualization, without modification.