医学图像体绘制中的快速三线性插值算法

目的 三线性插值是医学图像体绘制中的基本运算单元，每次采样后都需要进行，因此高效快速的采样计算是提高体绘制成像速度的重要途径之一，特别是在微型机上实现医学图像的体绘制。本文目的是提出了一种全新的快速三线性插值算法。方法 根据体数据中单个体素的8个顶点的数值分布，把全部体素分成128类，并用一个字节M中的各位来表示，插值计算时可根据每种体素的M值来选择相应的插值计算公式，从而极大地减少了插值计算的总量。此外，通过分类时的阈值设定，还可以灵活地改变三线性插值的运算总量。结果 针对体数据的特性，提出了一种高效灵活的三线性插值算法。结论 与其他快速三线性插值算法比较，该算法不仅能够显著减少计算量，还可以根据对结果图像的精度要求，通过设定的阈值对体数据的单个体素进行分类，灵活地调整三线性插值的运算总量，提高体绘制的速度。

A Fast Tri-linear Interpolation Algorithm for Volume Rendering of Medical Images

Objective Tri linear interpolation is a basic operation in volume rendering for medical images and is executed voxel by voxel to gain the value of each resampled point. The optimization of tri linear interpolation is an important approach of improving the speed of volume rendering, in particular, for the realization of volume rendering for medical images in a personal computer. Method According to the distribution of the values of 8 vertexes of each voxel, all voxels are classified into 128 types. A byte value M is used to denote the classified results. Different equation of interpolation is chosen in accordance with the value of M for each voxel when the operations of interpolation are executed, which can reduce the total operations of interpolation remarkably. Furthermore, by changing the threshold, which is used to determine the type of interpolation of voxel, the total operations of interpolation can be adjusted flexibly. Result According to the property of volume data, an effective and flexible tri linear interpolation algorithm was proposed. Conclusion Compared with other fast tri linear interpolation algorithms, the proposed algorithm has an additional advantage which allows the user to adjust the total operations of interpolation by changing the threshold of classified voxel to the accuracy of resulting image, and speeds up the volume rendering.