Transfer functions composition for volume data mining

Abstract

Analytic, simulated, scanned datasets were used to evaluate two representative data-centric methods for designing transfer functions (TFs), which are a key factor determining the quality of volume rendered images. They map the physical fields of a given volume dataset to the optical properties, such as color and opacity. Designing proper TFs is difficult because they depend on both the context of the target volume dataset and the purpose of the visual exploration. A system called "Ivory" was developed to assist in the design of TFs, With it, a better TF can be composed so as to inherit different features from two original TFs.