A Two-Stage Feature Extraction Approach for ECG Signals

This paper investigate various techniques of extracting features from the electrocardiogram (ECG) signal in order to analyze the ECG signals to detect the heart disease. Feature extraction, is a one of the widespread process of decompose the ECG data. This paper introduce a two-stage feature extraction approach to extract features from ECG signals for different types of arrhythmias. Firstly, Modified Pan-Tomkins Algorithm (MPTA) is implemented to remove noise and extract nine features. Then the proposed Improved Feature Extraction Algorithm (IFEA) is applied to extract additionally ten different features from the ECG signal. The MIT-BIH arrhythmia database have been used to test the proposed approach. It is obvious from the results that the proposed approach shows a high classification in terms of the following four statistical measures: Accuracy (Ac) 98.37%, Recall 48.29%, Precision 43.91%, F Measure 45.31%, and Specificity (Sp) 93.30%, respectively.