Dynamic time warping based arrhythmia detection using photoplethysmography signals

Photoplethysmography (PPG) based methods have gained popularity in recent times for arrhythmia detection. However, limited research has been carried out for multiple arrhythmia detection using PPG signals. Dynamic time warping (DTW) is a widely used time series technique for the comparison of speech and word recognition. However, the use of the DTW technique for arrhythmia detection using PPG signals is unexplored. In this research work, DTW is utilized to extract automated generated warping features. A feed-forward artificial neural network (ANN) has been used to classify the arrhythmia among four arrhythmia classes. The evaluation of the results has been carried out on 670 PPG signals of 8-s duration available on the PhysioNet MIMIC-II public database. The proposed model obtains an accuracy, sensitivity, specificity, F1 score, and precision of 95.97%, 97%, 97%, 96%, and 96%, respectively. The results show that the proposed approach has been able to detect multiple types of arrhythmias with significant performance.