Automated Detection of Myocardial Infarction and Heart Conduction Disorders Based on Feature Selection and a Deep...

Area of Science:

• Cardiology
• Biomedical Engineering
• Machine Learning

Background:

• Electrocardiograms (ECGs) are crucial for diagnosing heart conditions.
• Analyzing long-term ECGs is time-consuming and challenging for cardiologists.
• Automated tools are needed for efficient detection of critical cardiac events like myocardial infarction (MI) and conduction disorders (CDs).

Purpose of the Study:

• To propose a novel deep learning approach for detecting MI and CDs using large-scale PTB-XL ECG data.
• To address challenges associated with existing deep learning methods, including data variability and computational expense.
• To develop a computationally efficient and interpretable diagnostic tool for cardiac arrhythmias.

Main Methods:

• A new deep learning model was developed for feature extraction from ECG signals.
• A custom activation function was proposed to enhance model convergence.
• Extracted deep features were classified using a Support Vector Machine (SVM).
• The approach considered data filtering and addressed challenges from diverse datasets.

Main Results:

• The proposed method achieved a high overall accuracy of 99.20% for MI and CD detection.
• The combination of Convolutional Neural Networks (CNNs) for feature extraction and SVM for classification proved effective.
• The custom activation function demonstrated fast convergence properties.

Conclusions:

• The developed deep learning approach offers a promising and accurate solution for automated detection of MI and CDs.
• The method overcomes limitations of traditional deep learning models in ECG analysis.
• This work contributes to the advancement of computer-aided diagnosis in cardiology.