Analysis of Computer Aided Identification System for ECG Characteristic Points

Digital signal processing and data analysis are very often used methods in a biomedical engineering research. In this work, the descriptions of two detection algorithms for ECG characteristic points are enclosed. The detection algorithms presented in this work are based on Pan and Tompkins’ algorithm and wavelet transform for signal de-noising and detection of QRS complexes. In the first approach, efficient designed filters are focused on removing supply network 50 Hz frequency and baseline drift due to breathing. A special digital bandpass filter reduces false detection caused by the various types of interference present in ECG signals. The next process after filtering is differentiation followed by squaring, and then integration. The integrated signal is detected by thresholding for QRS complex. P wave and T wave detection are performed by using detected QRS complexes. MATLAB program is developed for the characteristic points’ detection. The algorithm for peak detection case is modified and it is applied to show ECG characteristic points. Wavelet Transform (WT) method is used for peak detection in this work. Wavelet based detection algorithms for one-dimensional signals are presented along with the results of detection ECG data. Firstly, ECG signals are decomposed by the Discrete Wavelet Transform (DWT). The decomposed signals are detected by thresholding for QRS complex. Detection of the QRS complex is the most important task in automatic ECG signal analysis. Finally, P wave and T wave detection are performed by using detected QRS complexes. Different types of algorithms are applied and evaluated their performance with sensitivity (Se), positive predictive (+P).