论文部分内容阅读
本文阐述一种动态心电快速分析的软件实现方法。心电快速分析任务划分为两个具有因果联系的子任务,由专门设计的双Intel 8088 CPU系统实现流水线处理。作者提出的差分阈值跳跃算法使R波检测速度成倍提高;用于QRS波模式分类的树分类器具有分类速度快、区分模式多等特点。分析系统具有基本的学习和适应功能,克服了心电信号的个体差异性和时间变异性问题,实验结果表明,分析速度达到实时速度的240倍以上,QRS波的分类正确率达95％以上。
This article describes a dynamic ECG rapid analysis software implementation. ECG rapid analysis task is divided into two sub-tasks with causal links, designed by the dual Intel 8088 CPU system to achieve pipeline processing. The differential threshold jump algorithm proposed by the author doubled the R-wave detection speed. The tree classifier used for the QRS wave pattern classification has the characteristics of fast classification and multiple discriminating modes. The analysis system has the basic functions of learning and adapting, overcoming the individual difference and time variability of ECG signals. The experimental results show that the analysis speed reaches over 240 times of the real-time speed and the correct classification rate of QRS wave is more than 95%.