Multi-dimensional transfer functions design based on boundary model

Direct volume rendering is an important method for visualizing complex dataset. However, many information is lost in the process of generating 2D data from 3D data. Since the lack of prior information of volume data, designing transfer function becomes a time-consuming and unintuitive work. Most direct volume renderings employ one-dimensional transfer functions, which assign color and opacity to the volume base on the single scalar quantity, but their capability for extracting specific material boundaries is not satisfied. In this paper, we describe the application of multi-dimensional transfer functions to scalar data. The data value´s first and second directional derivatives along the gradient direction help us find the exact boundaries of interested structures. Through analyzing the boundary model, we can quickly and easily design our transfer functions. Experimental results show that this approach can solve the locating of interested structures, and the rendering results completely meet clinical needs.