A Novel Pitch-Frequency-Based ECG Signal Classification Approach for Abnormality Detection

Electrocardiogram (ECG) has a significant role for measuring the electric activity of the heart to discover heart diseases. Accurate classification of the ECG signals is used to detect the heart abnormalities. The present work is an efficient approach for the classification of normal and abnormal ECG signals based on pitch frequency estimation of these signals. Two time-domain methods, namely the auto-correlation function (ACF), and average magnitude difference function (AMDF) are used for pitch detection from ECG signals. The receiver operating characteristic (ROC) curve is used to measure the accuracy of the proposed method for ECG signal classification. The results report 100% classification accuracy of the ECG signals.