An intelligent diagnosis method for cardiovascular diseases based on the CNN-CBAM-GRU model

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Medicine
• Cardiology

Background:

• Early diagnosis of cardiovascular diseases (CVDs) is critical for patient outcomes.
• Electrocardiogram (ECG) signal classification is challenging due to complex signal characteristics.
• Automated CVD diagnosis requires robust and efficient ECG analysis methods.

Purpose of the Study:

• To develop and evaluate a novel deep learning model for enhanced ECG signal classification.
• To improve the accuracy and efficiency of automated cardiovascular disease diagnosis using ECG data.
• To provide a comprehensive performance evaluation of the proposed model on public ECG datasets.

Main Methods:

• Proposed a hybrid deep learning model integrating Convolutional Neural Networks (CNN), Convolutional Block Attention Module (CBAM), and Gated Recurrent Units (GRU).
• Evaluated the CNN-CBAM-GRU model on two benchmark ECG datasets: MIT-BIH and PTB-XL.
• Conducted a comprehensive performance assessment using accuracy, precision, recall, sensitivity, and F1-score for five-class classification.

Main Results:

• The proposed CNN-CBAM-GRU model achieved high classification performance on both datasets.
• Achieved 98.17% accuracy and 98.91% F1-score on the MIT-BIH dataset.
• Achieved 99.21% accuracy and 99.47% F1-score on the PTB-XL dataset, with 2.45 million parameters.

Conclusions:

• The CNN-CBAM-GRU model offers a practical and robust solution for intelligent ECG classification.
• The model demonstrates a strong balance between predictive performance and computational efficiency.
• This approach facilitates automated cardiovascular disease diagnosis through advanced ECG analysis.