A Derivative-Based Approach for QT-Segment Feature Extraction in Digitized ECG Record

Automatic electrocardiogram analysis using different signal processing technique is one of the prominent areas of research in biomedical engineering. This paper illustrates a simple approach for time plane feature extraction in the QT segment. At first, the R peaks are accurately determined using a simple derivative-based approach and hence heart rate is calculated. Next, the baseline points of all cardiac cycles in the dataset are determined in the TP segment and the baseline modulation in the signal is eliminated by an empirical formula. Finally, the characteristic points Q, Q-offset, S, S-offset, T-onset, T and T-offset are calculated for all cycles using a magnitude and slope threshold based method. Hence, QRS width, ST segment, QT segment and (QT)_c are determined. ECG data for 30 second interval from MIT-PTB diagnostic database is used for testing the algorithm.