Designing Feature-based Transfer Functions for Volumetric Rendering of Gaseous Objects

Abstract

This paper focuses on the rendering of fluids represented by volumetric datasets. Generally, volume data can be visualized by a volume rendering technique. The result depends on the transfer function that converts volume data into colored volume. However, it is often difficult to create the desired images by manually adjusting the transfer function. We propose an intuitive system that allows the user to directly design the appearance of the volume by specifying the colors of the volume with a set of control points. The system solves an inverse problem to determine the transfer function such that the colors of the rendered image become the same as those of the control points. In our method, the transfer function is a multi-dimensional function of features computed from the input volume dataset. Our method represents the transfer function as a linear combination of radial basis functions to solve the inverse problem.