论文部分内容阅读
目的探索利用心率序列检测心律失常的方法。方法由MIT-BIH心律失常数据库的注释文件得到R-R间期序列及心拍标注信息。将心拍标注中的窦性心拍、束支传导阻滞、各类融合心拍、起搏心拍、未分类心拍定义为“正常心拍”;将早搏及逸搏定义为“非正常心拍”。对原始R-R间期序列依次进行中值滤波与均值滤波,得到平滑R-R间期序列;原始R-R间期与平滑R-R间期的差值过大者为“离群”R-R间期,相应的心拍归为“非正常心拍”。结果 MIT-BIH心律失常数据库全部48例记录中,有35例记录的窦性心拍数量占总心拍数量的一半以上。这35例记录“非正常心拍”检测的敏感度为88.2%,阳预测率为91.3%;对其中不含房颤或房扑的28例记录“非正常心拍”检测的敏感度为94.2%,阳预测率为96.4%。结论利用心率序列可以识别多数心室期前收缩与心房期前收缩,从而降低人工判读心律失常的工作量。
Objective To explore the method of using heart rate sequence to detect arrhythmia. Methods R-R interphase and heart beat annotation information were obtained from the annotation file of MIT-BIH arrhythmia database. The heart beat beating in the sinus heartbeat, bundle branch block, all kinds of fusion heartbeat, pacing heartbeat, unsorted heartbeat is defined as “normal heartbeat”; the premature beats and escape is defined as “abnormal heartbeat” . The original RR interval sequence is followed by median filtering and mean filtering to obtain a smooth RR interval sequence. The difference between the original RR interval and the smoothed RR interval is the “outlier” RR interval, and the corresponding Heart beat classified as “abnormal heart beat ”. Results Of the 48 records of MIT-BIH arrhythmia database, 35 recorded sinus beats accounted for more than half of the total number of beats. The sensitivity of the “abnormal heartbeat” test was 88.2% and the positive predictive rate was 91.3% in the 35 cases. The sensitivity of the “abnormal heartbeat” test was recorded in 28 cases without atrial fibrillation or atrial flutter 94.2%, positive predictive rate was 96.4%. Conclusion The use of heart rate sequences can identify most pre-ventricular systolic and atrial premature contractions, thereby reducing the manual interpretation of arrhythmia workload.