摘要
成像速度是影响虚拟内窥镜应用的关键问题之一。为了提高成像速度,本文给出一种基于Intel超线程技术(hyper-threading,HT)的虚拟内窥成像算法,在对传统的VE算法进行并行化处理的基础上,用OpenMP语言编程,使之能在支持HT技术的P4/3.06G的Pentium处理器上实现双线程并行计算,实现了对虚拟内窥镜系统优化的目的。与传统的虚拟内窥镜算法比较速度提高三倍多,而且对图像质量没有任何影响,具有实际应用价值。论文提供的算法还可应用于其他需要并行处理的场合。
Rendering speed is a key problem in the Virtual Endoscope (VE). In order to accelerate its rendering process, this paper presents a Vitural Endoscope algorithm by using Intel Hyper Threading (HT) technology. Based on the parallelization of traditional VE algorithm, we implemented its two-thread parallel computation on the P4/3.06G Pentium processor, which supports the HT technique, with OpenMP language. The algorithm is over three times faster than the traditional VE algorithm without any bad impact on the image quality so it is worthy in practicality. The programming idea mentioned here can also be used for other parallel applications.
出处
《北京生物医学工程》
2006年第3期248-251,284,共5页
Beijing Biomedical Engineering
基金
安徽省自然科学基金(01042202)资助
关键词
虚拟内窥镜
超线程技术
光线投影
并行算法
virtual endoscope hyper-threading technology ray casting parallel algorithm