A Novel Diagnostic Algorithm for Heart Disease in ECG Monitoring System

Abstract

Healthcare Internet of Things (HIoT) can connect mobile and wearable devices in the medical field, making disease monitoring and diagnosis possible anytime and anywhere. Most of these mobile and wearable devices can collect physiological signals of the human body in real time. Among them, ECG signal as a non-invasively collected signal that can effectively reflect the physiological changes of the heart plays a vital role in clinical and HIoT. We firstly propose a practical ECG monitoring system based on Humeds Portable ECG Monitor. Secondly, based on wavelet transform (WT) and deep convolutional neural network (DCNN), we propose a new algorithm suitable for the diagnosis of atrial fibrillation (AF) and arrhythmia. The sensitivity of AF is 0.978 and the accuracy rate of arrhythmia diagnosis is 0.991. Thirdly, we collected ECG data from 17 volunteers and verified the AF algorithm, the final average accuracy is 0.852. The ECG monitoring system designed in this paper can be used as a complete and effective application of HIoT. The algorithm designed in this paper is not only applicable to the ECG monitoring system proposed but also can be integrated as a potential algorithm in other ECG mobile and wearable devices.