Multi-ECGNet for ECG Arrythmia Multi-Label Classification

Junxian Cai,Weiwei Sun,Ilsun You,Jianfeng Guan

doi:10.1109/access.2020.3001284

Abstract

With the development of various deep learning algorithms, the importance and potential of AI + medical treatment are increasingly prominent. Electrocardiogram (ECG) as a common auxiliary diagnostic index of heart diseases, has been widely applied in the pre-screening and physical examination of heart diseases due to its low price and non-invasive characteristics. Currently, the multi-lead ECG equipments have been used in the clinic, and some of them have the automatic analysis and diagnosis functions. However, the automatic analysis is not accurate enough for the discrimination of abnormal events of ECG, which needs to be further checked by doctors. We therefore develop a deep-learning-based approach for multi-label classification of ECG named Multi-ECGNet, which can effectively identify patients with multiple heart diseases at the same time. The experimental results show that the performance of our methods can get a high score of 0.863 (micro-F1-score) in classifying 55 kinds of arrythmias, which is beyond the level of ordinary human experts.

Highlights

Electrocardiography is the process of producing an Electrocardiogram (ECG) which is a graph of voltage versus time of the electrical activity of the heart [1] using electrodes placed on the skin
In order to realize efficient, high-precision and highly automated end-to-end ECG detection, this paper proposes a Multi-ECGNet model based on one-dimensional convolution Resnet, depthwise separable convolution, and Squeeze-andExcitation (SE) Module, which can be used for multi-label classification detection of ECG data
We propose a Multi-ECGNet network architecture by incorporating ResNet, depthwise separable convolution, SE Module design and weighted binary crossentropy improved based on focal loss as a multi-label classification loss

Summary

Introduction

Electrocardiography is the process of producing an Electrocardiogram (ECG) which is a graph of voltage versus time of the electrical activity of the heart [1] using electrodes placed on the skin. ECG is the simplest and most efficient way to diagnose heart-related diseases and has been used in medical practice since 1903 [2]. After more than 100 years of development, ECG has played an important role in medicine, and many related advanced and easy-to-use clinical devices have been developed. How to use AI especial deep learning methods to carry out efficient and accurate medical diagnosis and free the manpower input of doctors is still a topic worthy of in-depth study. The method of measuring the ECG is to place the electrodes in different parts of the human body and connect them to the positive and negative electrodes of

Methods

Results

Conclusion