Sheng-Ta Hsieh and Chun-Ling Lin
Atrial Fibrillation (AFib) is the most common cardiac arrhythmia, increasing in prevalence with age. AFib is often associated with structural heart disease and a substantial proportion of AFib patients lead to the significant morbidity, mortality, and cost. Thus, AFib is the most prevalent and costly health problems in the world and a major global healthcare challenge. This study presents a beat-to-beat AFib detection system to provide a healthcare system for AFib patients. For real-time Electrocardiogram (ECG) signals, the beat-to-beat AFib detection system consists of two methods in this study: an improved ECG R peak detection method and a beat-to-beat Gaussian voting AFib method. The improved R peak detection method proposes two different optimization algorithms to replace the knowledge-based theory in previous R peak detection method that consists of three stages: band-pass filter, interesting blocks and threshold. The beat-to-beat Gaussian voting AFib method extracts features based on the R-R intervals to identify the possibility of AFib. Based the R-R intervals, the heart rate can be estimated, and the system can detect the tachycardia and bradycardia in addition. The results using the MIT-BIH database show that the proposed R peak detection method can detect beats with 99.9984% accuracy in testing data. Clinical testing reveals that the proposed beat-to-beat Gaussian voting AFib method is about 94.72% accurate and 98.11% sensitivity for 6 normal subjects and 6 AFib patients.
分享此文章
English 
Spanish 
Russian 
German 
French 
Japanese 
Portuguese 
Hindi 