Real Time P, QRS and T Wave Detection by QRS Matched Filter Method
ECG signals have always been a key concern for heart disease analysis and heart monitoring system. That’s why for a very long time, people are to find newer and easier processes for retrieving information from ECG. And still now some of the methods are quite accurate and implemented successfully all over the world, but still now many people are trying to find a more robust and simple method. My intention was to derive a simpler real time detection of P, QRS and T waves. For this purpose, I developed such a model which uses mostly peak thresholding approaches for detecting these waves. But the novelty of this model is that in almost all approaches people tried to filter ECG signals to avoid baseline shifts, removal of noises. I tried to find such a filter which can make QRS peaks completely visible along with removing noises and baseline shifts. I used to address this filter as QRS matched filter. Using this filter QRS peaks were successfully captured and then after further processing like cross correlations, P and T waves were also successfully detected. This model was tested on both AVEC 2016 ECG signal database for emotion recognition and MIT-BIH Arrhythmia database.
Keywords – QRS Complex, QRS Matched Filter, Half Wave Rectification, Signal Clipping, Bisection, Triangular Filter, and Cross-Correlation.