Optimizing Volume Rendering Based on Field Topology Analysis

点击翻译按钮获取中文摘要

Volume visualization has served as an indispensable methodology in various disciplines for exploring the inner structures and complex behavior of volumetric objects embedded in large-scale sampled or simulated 3D datasets. Although the advent of hardware-based acceleration mechanisms enables the user to visualize such volumes interactively, the rapid increase in their data size makes it difficult to adjust transfer functions sufficiently with repeated evaluation of resulting images. In this study, we take advantage of a 3D field topology analysis to optimize the transfer functions aiming at volume data mining. The conventional Reeb graph-based approach to describing the topological features of 3D surfaces is extended to capture the topological skeleton of a volumetric field. Based on the analysis results, which are represented in the form of hyper Reeb graphs, we proposed two principles to design appropriate color/opacity transfer functions for direct volume rendering. A feasibility study of the proposed methodology is performed with analytic trivariate functions, a large-scale 4D simulated dataset from atomic collision research, and a CT scanned tooth dataset.