Spindle Autoencoder-CNN hybrid model for cardiac arrhythmia classification.

Area of Science:

• Biomedical Engineering
• Cardiology
• Artificial Intelligence in Medicine

Background:

• Cardiac arrhythmias are irregular heart rhythms affecting blood circulation, commonly diagnosed via electrocardiograms (ECGs).
• ECG analysis is crucial for diagnosing heart rhythm disorders due to its reliability and cost-effectiveness.
• Automated arrhythmia detection systems are vital for enhancing diagnostic efficiency in clinical practice.

Purpose of the Study:

• To introduce a novel deep learning framework for classifying cardiac arrhythmias using ECG signals.
• To integrate a Modified Spindle Autoencoder (MSCAE) for feature extraction with a Convolutional Neural Network (CNN) for classification.
• To evaluate the proposed MSCAE-CNN model's performance on a standard arrhythmia database.

Main Methods:

• ECG signals from the MIT-BIH Arrhythmia Database were preprocessed and segmented into individual heartbeats.
• A Modified Spindle Autoencoder (MSCAE) was employed for deep feature representation extraction from ECG segments.
• A Convolutional Neural Network (CNN) was utilized to capture spatial relationships in the extracted features for arrhythmia classification.

Main Results:

• The integrated MSCAE-CNN model achieved a high classification accuracy of 98.78%.
• The proposed deep learning framework demonstrated superior performance compared to existing arrhythmia classification methods.
• Feature extraction using MSCAE effectively captured complex representations from ECG signals for improved classification.

Conclusions:

• The MSCAE-CNN framework offers a promising approach for rapid and accurate ECG-based arrhythmia detection.
• This model has significant clinical potential to aid medical decision-making in diagnosing heart rhythm disorders.
• Deep learning integration enhances the reliability and efficiency of automated cardiac arrhythmia diagnosis.