COVID-19 disease diagnosis from paper-based ECG trace image data using a novel convolutional neural network model

Area of Science:

• Cardiology
• Artificial Intelligence
• Infectious Diseases

Background:

• COVID-19 impacts the cardiovascular system, necessitating improved diagnostic tools beyond RT-PCR and imaging.
• Current diagnostic methods for COVID-19 have limitations in sensitivity and cost.
• Early detection of COVID-19's cardiovascular effects via electrocardiograms (ECG) is crucial.

Purpose of the Study:

• To develop and validate a novel deep Convolutional Neural Network (CNN) model for diagnosing COVID-19 using ECG trace images.
• To assess the model's ability to detect cardiovascular abnormalities linked to COVID-19.

Main Methods:

• A deep CNN model was designed to analyze ECG trace images derived from COVID-19 patients.
• The model was trained and tested on ECG data to classify various conditions, including COVID-19, normal heartbeats, and myocardial infarction.

Main Results:

• Achieved high binary classification accuracies: 98.57% (COVID-19 vs. Normal), 93.20% (COVID-19 vs. Abnormal Heartbeats), and 96.74% (COVID-19 vs. Myocardial Infarction).
• Demonstrated strong multi-classification performance: 86.55% and 83.05% for complex diagnostic tasks.
• Attained excellent Area Under the Curve (AUC) values, indicating robust diagnostic capability.

Conclusions:

• The proposed ECG-based deep learning model shows significant potential for rapid and accurate COVID-19 diagnosis.
• This approach can expedite patient diagnosis and treatment, optimize clinician workflow, and aid in pandemic management.