肿瘤图像特征统计中多线程技术应用研究

分析了肿瘤图像分析诊断医疗软件中影响速度的瓶颈模块,并应用多线程技术对瓶颈模块进行改造。在单核CPU和多核CPU平台上比较了不同线程数下运行速度、线程效率、统计误差和线程数的关系。最后对上述指标的平衡进行了讨论,在不影响辅助诊断效果的前提下给出适合的多线程改造方案。

A Research on Multi-threading Technology and Its Application in Statistics of Tumor Image Feature

The paper analyzes the bottleneck module of the cancer diagnostic medical image analysis software, which influences the running speed, and reforms the bottleneck module by employing the multi-threading technology. Then it compares the running speed, thread efficiency and statistical errors run on different threads with different number and their relationship among the numbers of threads. Finally the proportion of the various indicators is discussed, and a reform program for multi-threading technology is given without affecting diagnosis results.