Identification of QRS Segments of Electrocardiogram signals using Feature Extraction

For the intensive need of arrhythmia identification of patients it is necessary to explore the feature segments of electrical representation of heart signals i.e. ECG signals. A feature extraction algorithm is presented in this paper for QRS segment detection, which is based upon the amplitude and derivative characteristics of QRS complex segment. The performance of proposed algorithm is evaluated on the parameters of sensitivity (Se %) and positive predictivity (+P %) and compared with Pan Tompkins and GR algorithm for clinical database. Sensitivity (Se %) of proposed algorithm is calculated as 99.71% which is 0.46 % higher than Pan Tompkins and 0.02% higher than GR algorithm. The overall positive predictivity (+P %) of proposed algorithm is 99.69% which is 0.58% higher than Pan Tompkins and 0.2% higher than GR algorithm. The proposed algorithm performs superior as a result from parametric improvements for the MIT-BIH arrhythmia database.