One of the core problems in direct volume rendering is the specification of a transfer function that shows the important information of the dataset. The goal of this project to develop methods that automatically or semi-automatically identify interesting features in the data and create a transfer function that emphasizes them.
The main idea behind the method being developed is to automatically identify each material within the data set, and then create a transfer function which renders the boundaries between these materials by analyzing the data values and their derivatives. The result is a fast, informative rendering of the data with minimal user input.