Deep Learning Models for Arrhythmia Detection in IoT Healthcare Applications

https://doi.org/10.1016/j.compeleceng.2022.108011Get rights and content

Highlights

  • A novel deep learning models for arrhythmia detection in IoT healthcare applications are proposed.

  • It has the independence of the ECG signal quality, as operation is on the spectrograms of the ECG signals.

  • Ability to work on a single ECG lead, which reduces the complexity of operations.

  • Applicability of the proposed framework in the smart healthcare platform for real-time operation.

  • The proposed framework overcomes the overfitting problems and achieves high accuracy on several datasets.

Abstract

In this paper, novel convolutional neural network (CNN) and convolutional long short-term (ConvLSTM) deep learning models (DLMs) are presented for automatic detection of arrhythmia for IoT applications. The input ECG signals are represented in 2D format, and then the obtained images are fed into the proposed DLMs for classification. This helps to overcome most of the problems of the previous machine and deep learning models such as overfitting, and working on more than one lead of ECG signals. We use several publicly available datasets from PhysioNet such as MIT-BIH, PhysioNet 2016 and PhysioNet 2018 for model assessment. Overall accuracies of 97%, 98 %, 94 % and 91 % are obtained on spectrograms of MIT-BIH dataset, compressed MIT-BIH dataset, PhysioNet 2016 dataset, and PhysioNet 2018 dataset, respectively. Compared to the previous works, the proposed framework is more robust and efficient, especially in the case of noisy data.

Graphical abstract

Deep Learning Models for Arrhythmia Detection in IoT Healthcare Applications.

Image, graphical abstract
  1. Download : Download high-res image (100KB)
  2. Download : Download full-size image

Section snippets

INTRODUCTION

A heart rhythm disorder (arrhythmia) is any unusual, or out-of-time pulse in a person's heart rhythm. An arrhythmia may be accompanied by abnormal heart rhythms (a rapid pulse called tachyarrhythmia, or a slow pulse called slow-disorder-bradyarrhythmia) [1]. There are different types of arrhythmia:

  • Premature atrial contraction: This is an excessive additional contraction, caused by premature atrial contraction. This condition is simple and does not cause clinical complications [2].

  • Premature

Proposed framework for smart healthcare platform

The motivation for the design of the proposed framework is to use modern classification tools to incorporate arrhythmia detection into the existing e-health frameworks. In this section, an overview of the related research and some of the existing frameworks as well as challenges in such frameworks are investigated as a precursor to the adoption of the proposed framework for improved arrhythmia detection performance.

Smart wearable devices, such as smart bands and smartwatches, have become more

Proposed Strategy

The strategy of the proposed framework is to use a DLM with specified parameters to create a hierarchal designed file (hdf). The purpose of obtaining this file is to burn it into a hardware facility, such as raspberry pi or Arduino for real-life applications. In addition, it can be involved in a mobile application for smartphones or smart watches. Fig. 2 shows the data flow and architecture of the proposed strategy. As shown in Fig. 2, the proposed strategy consists of three main phases as

Simulation Results

The proposed DLMs are implemented on different datasets including MIT-BIH, PhysioNet 2016 and PhysioNet 2018. In addition, these DLMs are carried out at different scenarios (binary classes and Multiole classes). This variety of datasets and scenarios provides an accurate and fair evaluation of the proposed models. Furthermore, the strategy of the simulation experiments is to select an optimum model with a certain number of epochs. This optimum DLM is evaluated by the accuracy of detection.

Result Discussion

From the previous results, the proposed DLMs reveal high performance in the detection of arrhythmia. The strategy of evaluation of the proposed models is based on multi-class evaluation. In this strategy, the proposed models are carried out on different datasets with several scenarios. The proposed DLMs are carried out on the MIT-BIH with both spectrogram and compressed forms in addition to PhysioNet 2016 and PhysioNet 2018. These datasets are provided in both two-class and multi-classes

Conclusion

The main objective of this study was to develop a new model for automatic detection of arrhythmia. The main contribution is the design of novel CNN and ConvLSTM-based DLMs to classify the input ECG signals to normal or arrhythmia. We worked on three publicly available datasets from PhysioNet: MIT-BIH database, PhysioNet 2016 and PhysioNet 2018. Our model achieved accuracies of 97%, 98%, 94% and 91% on spectrograms of MIT-BIH dataset, compressed MIT-BIH dataset, PhysioNet 2016, PhysioNet 2018

Declaration of Competing Interest

The author declares that no conflict of interest exists and if accepted, the article will not be published elsewhere in the same form, in any language, without the written consent of the publisher.

Acknowledgments

This work was supported by the EIAS Data Science Lab, College of Computer and Information Sciences, Prince Sultan University, Riyadh, Saudi Arabia.

References (30)

  • R. Pranata

    Frequent premature atrial complex: A neglected marker of adverse cardiovascular events

    Int. J. Cardiovasc. Acad.

    (2020)
  • E.J. da S. Luz et al.

    ECG-based heartbeat classification for arrhythmia detection: A survey

    Comput. Methods Programs Biomed.

    (2016)
  • H. Sohal et al.

    Interpretation of cardio vascular diseases using electrocardiogram: A study

  • A.M. Shaker et al.

    Generalization of Convolutional Neural Networks for ECG Classification Using Generative Adversarial Networks

    IEEE Access

    (2020)
  • B.B. Gupta et al.

    Blockchain-assisted secure fine-grained searchable encryption for a cloud-based healthcare cyber-physical system

    IEEE/CAA Journal of Automatica Sinica

    (2021)
  • Cited by (45)

    • Application of Deep-Q learning in personalized health care Internet of Things ecosystem

      2023, Deep Learning in Personalized Healthcare and Decision Support
    • Automated Detection of Arrhythmia in ECG Signals using CNN

      2024, International Journal of Intelligent Systems and Applications in Engineering
    View all citing articles on Scopus
    View full text