Visualization of Irregularly Sampled Volumetric Data

可视化为理解复杂的数字数据提供了一种高效而强大的手段。该项目研究了采样体积数据的可视化方法，例如由计算流体动力学、层析成像和分子模拟产生的数据，特别是那些样本在空间中不规则分布的体积。可视化的两种常见方法是：1)提取感兴趣的特征(如等值面)；2)将While体积成像为连续的半透明介质。大多数现有的可视化方法都是针对规则的直线网格设计的。本项目将扩展和推广这些方法，以便以不太规则的格式显示数据，这些格式在许多应用程序中都可以找到。两种主要方法是：1)将现有算法扩展到不规则体积；2)将不规则体积重采样为分层规则体积。这项研究的另一个方面是仔细比较可视化某些方面的不同方法，特别是3D数据的重新采样(内插)、从数字数据到颜色和透明度等视觉参数的映射，以及高效的体遍历方法。

Visualization provides an efficient and powerful means of comprehending complex numeric data. This project investigates approaches to visualization of sampled volumetric data, such as that produced by computational fluid dynamics, tomography, and molecular simulations, particularly those volumes where the samples are irregularly distributed in space. Two common methods for visualization are: 1) to extract features of interest (such as isosurfaces); and 2) to image the while volume as a continuous semi-transparent medium. Most present approaches to visualization are designed for regular rectilinear grids. This project will extend and generalize these approaches to visualization of data in the less regular formats which are found in many applications. The two main approaches will be: 1) to extend existing algorithms to irregular volumes; and 2) to resample the irregular volumes to hierarchical regular volumes. Another aspect of this research is to carefully compare different approaches to certain aspects of visualization, particularly, the "resampling" (interpolation) of 3D data, the mapping from numeric data to visual parameters such as color and transparency, and methods of efficient volume traversal.