부정맥 판독을 위한 형태변형 및 간격변형 기반 특이심박 검출

Abstract

The abnormal heartbeat caused by the occurrence of arrhythmia greatly affects arrhythmia reading, so abnormal heartbeat detection is very important. In this paper, we propose an abnormal heartbeat detection method based on the shape and interval deformation of the QRS complex. The shape deformation is represented by the changes of amplitude and kurtosis in R-peak of QRS complex, and the interval deformation is represented by the change of interval between QRS complexes. The distribution of the amplitude and the kurtosis in R-peak and the interval between QRS complexes follow the normal distribution in the normal ECG signal, but do not in the arrhythmia ECG. On the basis of this, we propose an abnormal heartbeat detection method using the median and the median absolute deviation of the each features. For the performance evaluation of the proposed method, MIT-BIH ADB which is a representative arrhythmia database was used. Through experiments of the proposed method, the non-detection rate of abnormal heartbeat was 0% and the over-detection rate of normal heartbeat was 1.78% on average. Therefore, it was confirmed that heartbeat data for arrhythmia reading was minimized through the proposed method.