Recognition and classification of electrocardiogram signal based on convolutional neural network

Abstract

Nowadays, the prevalence of cardiovascular diseases is increasing, leading to a rise in mortality rates. The electrocardiogram (ECG) is a common way to track activity of heart, some cardiovascular diseases including arrhythmia can be diagnosed with the ECG. However, diagnosing cardiac arrhythmias still heavily relies on the subjective judgment of cardiovascular specialists based on ECG readings. This project proposes a model that utilizes machine learning algorithms for the classification and processing of ECG signals. After collecting relevant data and performing necessary data preprocessing steps, the Convolutional Neural Network (CNN) model is employed for recognizing and classifying. Experimental findings demonstrate that the proposed method performs excellently in classifying various common types of cardiac arrhythmias, yielding high accuracy of over 99%. The method exhibits fast processing speeds, making it suitable for real-time ECG analysis and monitoring. Therefore, the outcomes of this project hold significant clinical relevance in improving the diagnosis of cardiovascular diseases and enhancing patients' quality of life.