中文摘要：

为了在保证绘制图像质量的基础上将体绘制算法的绘制速度提高至实时,提出一种基于可编程图形加速硬件（GPU）的光线投射算法实现（GRC,GPU-based Ray Casting）。GRC在可编程GPU中进行重采样和分类,使用矩阵逆运算以降低重采样坐标的计算复杂度,使用后分类技术以降低算法的空间复杂度。实验表明：对于2563规模的体数据,GRC能够在保证图像质量的基础,以超过30fps的速度进行绘制。

英文摘要：

An implementation of ray casting on programmable GPU （Graphics Process Unit） was proposed for achieving real-time performance of volume rendering algorithms as well as high-quality images. GRC, short for GPU-based ray casting, takes the processes of resampling and classification in GPU. Inverse-matrix projection is used for reducing the computing complexity of resamples＇ coordinates. Post-classification is employed to control the required main memory and texture memory of ray casting. The experiments show that GRC can produce high-quality images at the rate of over 30 fps for 2563 volume datasets,