Vol. 15, no.4, 2023
РусскийEnglish
MEDICAL PHYSICS
TOWARD AN OPTIMAL WAVELET FILTER AND DECOMPOSITION LEVEL FOR NOISE ELIMINATION OF THE ECG SIGNAL
Anas Fouad Ahmed
Al-Iraqia University, Electrical Engineering Department, https://en.aliraqia.edu.iq/
Al Adhmia - Haiba Khaton, 6029, Baghdad, Iraq
E-mail: anasfuad33eng@yahoo.com
Ali Rasim Ibrahim
Alsalam University College, https://alsalam.edu.iq/en/
Baghdad - Sidiya, Iraq
E-mail: ali.r.ibrahim@alsalam.edu.iq
May Hatem Abood
Al-Iraqia University, Computer Engineering Department, https://en.aliraqia.edu.iq/
Al Adhmia - Haiba Khaton, 6029, Baghdad, Iraq
E-mail: may.hattim@gmail.com
Received March 28, 2023, peer-reviewed April 28, 2023, accepted May 05, 2023, published December 06, 2023.
Abstract: The denoising process represents one of the most important preprocessing steps for Electrocardiogram (ECG) signal processing and assists the specialist in making the right diagnosis for the patient. Five wavelet filters (WFs) closest in morphology to the pattern of the ECG signal were nominated, and their performances were analyzed at different noise, and number of decomposition (No. Dec) levels to determine the optimum, among them for noise reduction task. These Filters are Daubechies 4 (DB4), Daubechies 6 (DB6), Coiflet 4 (Coif4), Symlet 6 (Sym6) and Symlet 8 (Sym8). The results of the standard ECG signals (downloaded from MIT-BIH) revealed that the DB6 filter with four decomposition levels is optimal for removing noise of the ECG signal in terms of three metrics "Mean Square Error" (MSE), "Output Signal to Noise Ratio" (SNRo), and "Correlation Coefficient Index" (CCI). In addition, a simple and efficient threshold rule was adapted to be used in the proposed method. The suggested approach was successfully applied to reduce the noise of the ECG signals recorded using a simple proposed electronic circuit. Finally, the performance of the introduced scheme was compared with that of the standard ECG equipment, the "Biocare iE300", and the outcomes were very close.
Keywords: electrocardiogram, noise elimination, wavelet filter, thresholding
UDC 53.047:57(075.8)
RENSIT, 2023, 15(4):401-410e DOI: 10.17725/rensit.2023.15.401
Full-text electronic version of this article - web site http://en.rensit.ru/vypuski/article/520/15(4)401-410e.pdf