Automatic detection of arrhythmias from an ECG signal using an auto-encoder and SVM classifier

Area of Science:

• Cardiology
• Biomedical Engineering
• Artificial Intelligence in Healthcare

Background:

• Millions worldwide suffer from potentially life-threatening heart arrhythmias.
• Electrocardiograms (ECG) are crucial for diagnosing these abnormal heart rhythms.
• Automated detection systems can significantly aid in timely diagnosis and management.

Purpose of the Study:

• To develop and evaluate a computer-aided diagnosis (CAD) system for automatic classification of arrhythmias from ECG signals.
• To improve the accuracy and efficiency of arrhythmia detection using machine learning techniques.
• To classify normal sinus rhythms (NSR) and four types of arrhythmias: left bundle branch block (LBBB), right bundle branch block (RBBB), paced beat (PB), and premature ventricular contractions (PVC).

Main Methods:

• Utilized an auto-encoder convolutional network (ACN) model, based on 1D-CNN, for automatic feature extraction from raw ECG signals.
• Employed a support vector machine (SVM) classifier on ACN-extracted features for enhanced arrhythmia beat detection.
• Validated the system using a tenfold cross-validation strategy on the MIT-BIH arrhythmia database.

Main Results:

• The proposed system achieved an overall accuracy of 98.84% in classifying arrhythmias using the SVM classifier.
• Demonstrated superior performance compared to existing methods in the literature for arrhythmia detection.
• High sensitivity and precision were reported for the classification of NSR, LBBB, RBBB, PB, and PVC.

Conclusions:

• The developed CAD system effectively classifies various arrhythmias from ECG signals with high accuracy.
• The ACN and SVM integrated approach shows significant potential for clinical application in cardiology.
• This automated system can support further clinical studies and improve patient care for cardiac conditions.