一种双导联融合心电QRS波检测算法  

A QRS Complex Detection Algorithm of Two-lead ECG Signal

作　　者：朱晓君[1] 杨红官[1] 刘健[2] 

机构地区：[1]湖南大学物理与微电子科学学院,湖南长沙410082 [2]北京科技大学计算机与通信工程学院,北京100083

出　　处：《中国医学物理学杂志》2015年第1期68-71,85,共5页Chinese Journal of Medical Physics

基　　金：国家自然科学基金(61173149);北京市青年英才计划项目

摘　　要：目的:心电图的检测与分析是临床上诊断心血管疾病的主要依据,QRS波是心电信号中最重要的特征波,它的准确检测是心电信号自动分析的前提和基础。为了提高单导联检测QRS波的灵敏度和准确率,本文提出一种新的双导联融合心电QRS波检测的算法。方法:原始心电信号通过单导联预检波进行QRS波定位后,由双导联决策方法来决定采用单导联检波还是双导联融合检波。单导联检波直接采用第一或第二导联检测结果;双导联融合检波由双导联融合方法和导联判断规则判别。以窗时间为时间单位,不断更新双导联决策算法。本算法包括方差、幅值、模板匹配以及阈值比较等方法。结果:采用MIT-BIH心律失常数据库的48组两导联心电记录进行验证,统计得到平均灵敏度和准确率分别为99.87%、99.81%。其漏检数和误检数比第一导联分别降低了23.26%、18.27%,比第二导联分别降低了88.21%、95.11%。结论:本算法实时高效地提高单导联QRS波检测的灵敏度和准确率,且优于部分算法的检测结果,因而在心电信号自动分析中具有良好的应用前景和较高的实用价值。Objective DDetection and analysis of Electrocardiogram （ECG） are the major basis for clinical diagnose ofcardiovascular disease. As the most important feature wave of ECG signal, the accurate detection of QRS complex defining isan important prerequisite and foundation for automated ECG analysis. In order to improve the sensitivity and specificity of thesingle-lead QRS detection, a new algorithm of detecting two-lead QRS complex is proposed. Methods The original ECG isdetected by each lead with single-lead QRS detecting rule, and determined to do single-lead examination or two-lead fusionexamination by two-lead decision algorithm. Examination of single-lead adopts first-lead or second-lead detection, whilefusion examination of two-lead is decided by two-lead fusion method and lead judge rule. The two-lead decision algorithm isupdated constantly with the unit of time window. This algorithm concludes a few methods, such as variance, amplitude,template matching, threshold comparison, and so on. Results The algorithm is evaluated by 48 ECG records of two leads in theMIT-BIH arrhythrnia database, and the achieved scores indicate high performance： 99.87% sensitivity and 99.81% specificityfor QRS detection. And the false negative and false position in the algorithm are reduced by 23.26% and 18.27% respectivelycompared to the first lead, which is reduced by 88.21% and 95.11% respectively relative to the second lead. Conclusion Theexperiment shows that the algorithm can improve sensitivity and specificity of single-lead QRS detection in real-time and itoutperform some published results of other detection algorithms. Thus the algorithm has a good application potential and highpractical value in automated ECG analysis.

关 键 词：双导联 QRS波检测 决策方法 融合检波 模板匹配 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]