基于树莓派的实时心脏病辅助诊断

依据心电图特征参数R峰面积与S峰面积比，对正常、异常心电图进行快速二分类以提高心电分类效率和准确率。通过MIT-BIH数据库统计分析，正常心电图R峰面积与S峰面积比远大于异常心电图中的面积比值。在QRS波群检测中，改进QRS波群起止点检测算法，将移动窗口积分法、距离最大值法与斜率突变法3种方法加权平均结果作为QRS起止点最终检测结果，弥补单一算法的不足，提高起止点检测精确度，为面积比计算及后续心脏病辅助诊断提供可靠前提。基于低功耗、便携的树莓派平台实现算法，以树莓派Raspbian系统自带的Python为工具，构建了一个便携、实时的心电辅助诊断平台。

Real-time Auxiliary Diagnosis of Heart Disease Based on Raspberry Pi

In this paper， the area ratio of R peak and S peak in electrocardiogram （ECG） is presented to distinguish normal ECG and abnormal ECG quickly， and improve the efficiency and accuracy of ECG diagnosis. According to the statistics of MIT-BIH database data， the ratio of R peak area to S peak area in normal ECG is far greater than which in an abnormal ECG. In QRS complex detection， the algorithm of detecting the starting and ending point is improved. The result of weighted average of three methods including moving window integration， maximum distance method and slope mutation is used as the final detection result of the starting and ending point of QRS complex， which provides a reliable premise for the calculation of area ratio and the subsequent auxiliary diagnosis of heart disease. The implementation of the algorithm is based on the low power and portable raspberry Pi platform. The algorithm implementation tool is Python which is a system software in Raspbian system， a portable and real-time auxiliary ECG diagnosis platform is constructed.