Related Experiment Videos

Explainable hybrid deep learning framework with Grad-CAM for heartbeat-level arrhythmia classification

Sureshkumar Sundaramoorthy1, Govardhan Karunanidhi1

  • 1School of Electronics Engineering, Vellore Institute of Technology, Vellore, Tamilnadu, India.

Frontiers in Physiology
|August 18, 2026
PubMed

Insights

This study introduces an explainable deep learning model for accurate electrocardiogram (ECG) arrhythmia classification. The advanced framework achieves high performance, offering a scalable solution for real-time cardiac monitoring.

Area of Science:

  • Cardiology
  • Artificial Intelligence
  • Biomedical Engineering

Background:

  • Cardiac arrhythmia is a major cause of death globally.
  • Early detection of ECG arrhythmias is crucial for patient outcomes.
  • Current diagnostic methods require timely and accurate classification.

Purpose of the Study:

  • To develop an explainable hybrid deep learning framework for automated ECG cardiac arrhythmia classification.
  • To improve the accuracy and interpretability of arrhythmia detection.
  • To create a robust and scalable solution for clinical application.

Main Methods:

  • Integration of 1D-CNN for local morphological features and GRU for temporal dependencies.
  • Application of channel attention and Grad-CAM for enhanced feature extraction and interpretability.
  • Evaluation on three benchmark ECG datasets (MIT-BIH Arrhythmia, INCART, SVDB).

Main Results:

  • Achieved high classification accuracy (up to 99.73%) and macro-F1 scores (up to 97.21%).
  • Demonstrated performance improvement of 2%-5% over state-of-the-art methods.
  • Showcased effective integration of morphological and temporal feature analysis.

Conclusions:

  • The proposed framework offers a robust and scalable solution for ECG arrhythmia classification.
  • Explainability through Grad-CAM enhances clinical trust and understanding.
  • Computational efficiency and single-lead ECG use enable real-time deployment in diverse settings.

Related Concept Videos