ECG Heartbeat Classification Using Convolutional Neural Networks

Xuexiang Xu,Hongxing Liu

doi:10.1109/access.2020.2964749

Abstract

Electrocardiogram (ECG) data recorded by Holter monitors are extremely hard to analyze manually. Therefore, it is necessary to automatically analyze and categorize each heartbeat using a computer-aid method. Because convolutional neural networks (CNNs) can classify ECG signals automatically without trivial manual feature extractions, they have received extensive attention. However, it is anticipated that improving the existing CNN classifiers might provide better overall accuracy, sensitivity, positive predictivity, etc. In this study, we proposed a CNN based ECG heartbeat classification method. Based on the MIT-BIH arrhythmia database, our proposed method achieved a sensitivity of 99.2% and positive predictivity of 99.4% in VEB detection; a sensitivity of 97.5% and positive predictivity of 99.1% in SVEB detection; and an overall accuracy of 99.43%. Our proposed system can be directly implemented on wearable devices to monitor long-term ECG data.

Highlights

In cases with suspected arrhythmias, doctors often ask the subjects to wear a Holter to continuously record ECG data for 24 hours or longer
According to the Association for the Advancement of Medical Instrumentation (AAMI) [2], non-life-threatening arrhythmia signals can be divided into five categories: nonectopic beat (N), supra ventricular ectopic beat (SVEB, S), ventricular ectopic beat (VEB, V), fusion (F), and unknown (Q)
In this study, we designed a Holter data convolutional neural networks (CNNs) heartbeat classifier based on the MLII lead by using coupled-convolution layer structure and adopting the dropout mechanism

Summary

Introduction

In cases with suspected arrhythmias, doctors often ask the subjects to wear a Holter to continuously record ECG data for 24 hours or longer. Because the amount of ECG data recorded by the Holter is extremely large, it is necessary to analyze the recordings using a computer and categorize the type of each heartbeat automatically [1]. We focus on classifying these 5 types of arrhythmia heartbeat. With the wide use of convolutional neural networks (CNN) in many fields [8]–[10], CNN have become a popular option to automatically classify ECG signals recorded by the Holter [3]–[7]. The CNN classifier can directly input heartbeats without additional feature extraction and selection; it demonstrates competitiveness in classification accuracy [3], [11], [16]

Methods

Findings

Discussion

Conclusion