您的位置:首页  > 论文页面

基于GPU统一计算设备架构的Micro CT图像重建

发表时间:2010-04-30  浏览量:1908  下载量:721
全部作者: 程杰,罗守华
作者单位: 东南大学生物科学与医学工程学院
摘 要: 针对三维锥形束计算机断层扫描(computerized tomography,CT)图像重建运算量大、重建时间较长的问题,提出运用图形处理器(graphics processing unit,GPU)及统一计算设备架构(compute unified device architecture,CUDA)实现用于三维锥形束CT图像重建的FDK算法。与中央处理器(central processing unit,CPU)实现方法相比,新方法在滤波卷积阶段和耗时最多的反投影阶段都得到了很高的加速比。而与传统的图形应用程序接口(application programmers interface,API)加速方法相比,所提出的方法更易于实现、效率更高。
关 键 词: 计算机应用;图像重建;FDK算法;图形处理器;统一计算设备架构
Title: GPU and CUDA based on Micro CT reconstruction
Author: CHENG Jie, LUO Shouhua
Organization: School of Biological Science and Medical Engineering, Southeast University
Abstract: The three-dimensional cone-beam computerized tomography (CT) reconstruction is compute-intensive and time-consuming. This paper presents an innovative implementation of FDK algorithm, which solves the three-dimensional reconstruction task in cone-beam CT, based on graphics processing unit (GPU) using compute unified device architecture (CUDA). This implementation achieved a higher-speed reconstruction performance compared with the central processing unit (CPU) method in the filtering stage and back-projection stage, which is the most time-consuming part. Besides, comparing with the traditional accelerated method with graphics application programmers interface (API), the method is easier to implement and more efficient significantly.
Key words: computer application; image reconstruction; FDK algorithm; graphics processing unit; compute unified device architecture
发表期数: 2010年4月第8期
引用格式: 程杰,罗守华. 基于GPU统一计算设备架构的Micro CT图像重建[J]. 中国科技论文在线精品论文,2010,3(8):819-824.
 
0 评论数 0
暂无评论
友情链接