标题：A decision-making fusion method for accurately locating QRS complexes from the multiple QRS detectors
作者：Liu, Feifei ;Liu, Chengyu ;Jiang, Xinge ;Zhao, Lina ;Li, Jianqing ;Song, Chuanjie ;Wei, Shoushui
作者机构：[Liu, Feifei ;Liu, Chengyu ;Zhao, Lina ;Li, Jianqing ] School of Instrument Science and Engineering, Southeast University, Nanjing; 210096, China;[Jia 更多
会议名称：World Congress on Medical Physics and Biomedical Engineering, WC 2018
会议日期：3 June 2018 through 8 June 2018
关键词：Decision-making fusion; Electrocardiogram (ECG); QRS detection
摘要：QRS detection for electrocardiogram (ECG) signal plays a fundamental role in monitoring cardiovascular diseases. Lots of QRS detection algorithms exist and most of them are verified with high sensitivity and positive predictivity on the standard ECG databases. Recent progress in mobile ECG rises the challenge of accurate QRS detection for real-time dynamic ECG recordings since the variety of noises. In this study, a decision-making fusion method for accurately locating QRS complexes from the multiple QRS detectors were proposed. First, the ECG signals were detected by these nine detectors. Then, the voting fusion rule had been established that a heartbeat was determined when more than five detectors showed their detections in a time moving window respectively. And the mean value of the middle three detections’ positions in the window was served as a corrected heartbeat. Moreover, the comprehensive post processing technology was used to eliminate the false detection and to search the missed beats. The new proposed method was tested on high and poor signal quality ECG databases. For comparison, the best detection accuracy for the single algorithm was only 75.50% while the new proposed fusion method with 200 ms time moving window reported a detection accuracy of 80.43% for the poor-quality ECG signals. The proposed fusion method can significantly improve locating QRS complexes accuracy for the ECG signals with poor signal quality. Thus, it has a potential usefulness in the real-time dynamic ECG monitoring situations. © Springer Nature Singapore Pte Ltd. 2019.