Synapse
⌘+K
Synapse
PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
June 22, 2023The Journal of Engineering and Exact SciencesOpen Access

Classification of ECG signals using deep neural networks

View Full Paper
Ask AI
Bookmark
Share

Why the study?

ECG classification using deep learning has shown promising results, and reliable methods are needed to accurately diagnose and classify cardiac diseases using ECG data.

Comparison

GoogleNet vs AlexNet vs ResNet Deep-CNN models

Key result

The GoogleNet deep convolutional neural network model achieved an accuracy of 96% in classifying ECG signals into normal sinus rhythm, congestive heart failure, and arrhythmia.

Authors

MNMohamed NadourLCLakhmissi CherrounNHNadji Hadroug

Discussion

Loading...

Member takes

Overview

CNNs classify ECG signals at up to 96% accuracy; leaves open prospective clinical validation before practice change.

Structured PICO

P
Population
162 ECG recordings from Physionet databases, segmented into 900 records, used to train and test deep learning models for classifying normal sinus rhythm, congestive heart failure, and arrhythmia.
I
Intervention
Deep convolutional neural networks (GoogleNet, AlexNet, ResNet) using 2D Scalogram images obtained through continuous wavelet transform (CWT) of ECG signals.
O
Outcome
Accuracy, precision, recall, and F1 score in classifying ECG signals into arrhythmia, congestive heart failure, and normal sinus rhythm.

Deep learning models, particularly GoogleNet, can accurately classify ECG signals into normal sinus rhythm, arrhythmia, and congestive heart failure using continuous wavelet transform scalograms.

Limitations

  • Results are based on an assessment specific to the dataset used and may vary depending on the quality of the dataset and the training parameters chosen.
  • Results are based on an assessment specific to the dataset used and may vary depending on the quality of the dataset and the training parameters chosen
  • Lack of cross-validation on different datasets

Cite This Study

Nadour et al. (2023) studied Arrhythmia and Congestive Heart Failure (n=162). Deep Convolutional Neural Networks (GoogleNet, AlexNet, ResNet) was evaluated on Classification accuracy. The GoogleNet deep convolutional neural network model achieved an accuracy of 96% in classifying ECG signals into normal sinus rhythm, congestive heart failure, and arrhythmia.

synapsesocial.com/papers/6a916198963a1f481f33f74fhttps://doi.org/10.18540/jcecvl9iss5pp16041-01e
View Full Paper
Ask AI
Bookmark
Share