硬件加速的大数据量自适应体绘制

利用树形结构和纹理映射技术,在普通微机上实现对大数据量体数据的实时交互.依靠八叉树结构和显卡的硬件加速功能,将体数据划分为不同精度的数据块,打破了大数据显示时显存与内存间容量和带宽的限制,通过交互策略动态遍历该树,实现对大数据量体数据多精度的绘制.实验结果表明,文中方法在普通微机上可以大于10帧/s的速度交互操纵GB级以上的体数据.该方法可有效地降低体绘制对于硬件的需求,使得在较低配置下对其交互成为可能.

Hardware-Assisted Adaptive Volume Rendering for Mass Data

By taking advantage of tree structure and texture mapping technique,interactive volume rendering for mass data can be made on ordinary personal computers. Utilizing the octree algorithm with graphics hardware acceleration,the original volume data is divided into multi-resolution blocks,so the problem of the transfer capacity and bandwidth restricts between graphics memory and the main memory can be eliminated. Rendering of huge volume data in various resolution is achieved by dynamically traversing the node...