An ECG classifier using fir with a QRS response

Abstract EN

Recording the electrocardiography (ECG) signals is one of the main ways to diagnose heart disease cases through detecting the total or partial alterations in the P wave, QRS complex, or T wave of the signal.

this bioelectrical signal is used to detect and visualize any irregular changes in the heart's electrical activity.

the two fundamental steps for ECG signal analysis are identifying the QRS complex and providing more details on the disease under medical investigation.

the discrete-time response of a high-pass FIR filter is achieved by equating the time coefficients of the filter with a discretized version of the QRS complex part of a standard ECG signal.

the high-pass FIR filter helps reduce the interference of the 50 Hz power supply.

a matched filter is designed using the discrete-time response of the high pass FIR filter, and the resulting matched filter response simulates a complex part of the QRS of an average ECG signal.

when an ECG of a patient is applied as an input to the designed matching filter, the best fit correlated output coefficients of the patient ECG are obtained.

the filter output coefficients are applied to the artificial neural network (ANN) recognition system to verify and classify the ECG input to four categories which are(AF, CHF, QT, and PTB) .

the overall system can operate with an accuracy of 98%.