Low Cost, Portable ECG Monitoring and Alarming System Based on Deep Learning

Abstract

Electrocardiogram (ECG) has been the golden standard for the detection of cardiovascular disease for many years. Any electrical impulse disruption that causes the heart to the contract may lead to arrhythmia. Arrhythmia patients have no indications of having an arrhythmia, but a doctor may recognize arrhythmias in a routine test. Therefore, continuous wearable personal monitoring system plays a big role, and it's become popular day by day. This research focuses on designing and developing a method for predicting arrhythmia (atrial fibrillation) along with monitoring the ECG signals. To create an arrhythmia prediction model and an Android-based real-time ECG surveillance system, Long Short-Term Memories neural network, Recurrent Neural Network, TensorFlow and Keras library are applied here. Those deep learning models and algorithms help to achieve overall 97.57 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> accuracy on arrhythmia prediction. The system is being designed with Raspberry pi 3, Arduino UNO, AD8232 single lead ECG sensor, HC-05 Bluetooth, biomedical sensor pad and battery. This system will make easier for doctors to monitor the ECG of their patients outside the hospital and also help for remote ECG monitoring. The total components cost of this research work is around USD 58.