Cardiac Signal Intelligence: ECG Signals Classification Using Deep Learning

ECG (Electrocardiogram) is a non-invasive device to measure the electrical activity of the heart in the form of ECG signals. There are many components of ECG signals that detect arrhythmia from the heart. Arrhythmia is group of diseases related to human heart. There are basically two types of arrhythmias life threatening and not life threatening. Life threatening such as heart stock and heart attack. So the aim of this study is to identify life threatening arrhythmic beats from ECG signals to reduce the risk factors from humans and find the treatment of heart diseases on time. In this study a light weight Convolutional Neural Network (CNN) model is proposed to identify the more dangerous heart beats from ECG signals. The CNN model is used on the MIT-BIH arrhythmia database which is publicly available and classifies different types of heart beats. The heart beats in the database are classified into five types such as normal (N), premature ventricular (V), premature atrial (A), left and right bundle branch block (L and R). Among them, the V and A types of heart beats are more dangerous heart beats. The CNN model is used to classify these heart beats and good results have been obtained in terms of accuracy, precision, recall and F1-score as shown in the results section.