Detection of Ventricular Arrhythmia by using Heart rate variability signal and ECG beat image

Ventricular Arrhythmia (VA) such as Ventricular Tachycardia (VT) and Ventricular Fibrillation (VF) are the common type of arrhythmia in infants and children. Electrocardiogram (ECG) signal is used for the diagnosis of such type of cardiac abnormality. Manual ECG assessment is an error prone task because of vast difference in ECG morphology. Hence, a computer aided diagnosis (CAD) system for classification of cardiac abnormality can be useful in cardiac care units. Main goal of our study is to design a CAD system to classify ECG signals of VT and VF patients. This study proposed a technique for the process of cardiac scoring which is a significant diagnosis step in VA detection. ECG signals are used for the extraction of Heart Rate Variability (HRV) signals and ECG beat images. The features are computed from the HRV signals and ECG beat images using thirty different feature extraction methods. Skewness, Kurtosis, shape factor, fractal dimension, entropy, contrast and dissimilarity are the selected features used for classification process. The best classification accuracy has been achieved by the use of ensemble classifier. According to performance analysis, the system can be operated with 99.99% accuracy rate for the separation between healthy and VA persons.

Automated summary

The study aims to design a CAD system for classifying ECG signals of VT and VF patients, proposing a technique for cardiac scoring. Feature extraction methods are used for computing features from HRV signals and ECG beat images, with the best classification accuracy achieved by an ensemble classifier. The system can operate with a 99.99% accuracy rate for distinguishing between healthy individuals and those with VA.