ECG Abnormality Classification with Single Beat Analysis

Abnormal electrical activity of the human heart indicates the cardiac dysfunction. Electrocardiogram(ECG) is a non invasive technique used as a diagnostic tool to detect cardiac abnormalities. Automatic recognition of abnormal ECG beats aids in early detection of heart diseases. Irregularity and non-stationarity in the ECG signal imposes difficulties to clinicians for accurate diagnosis of heart diseases. Signal processing algorithms can reveal better information within the noisy ECG signal. This paper explores the ECG deniosing and feature extraction from single ECG beat to support the ECG abnormality classification. Binary classification is implemented using Support Vector Machine(SVM). This technique has been extended to multi-class classification for non linearly separable data effectively. In this work, classification accuracy upto 88% is achieved for selected input feature set. This work assesses the suitability of the feature set for multi-class classification of cardiac diseases.