Baseline Wander Elimination from Electrocardiogram Signals Using Dyadic Boundary Points-Based Empirical Wavelet Transform

Baseline wander (BW) refers to low-frequency fluctuations in the baseline of electrocardiogram (ECG) signals, which can obscure the underlying physiological information. It is important to address BW to obtain accurate and useful data in various applications. To expunge BW from the ECG we have used filter bank-based technique in this manuscript. The best filter bank for multiscale decomposition of the ECG to different sub-band (SBD) signals is introduced. It relies on dyadic boundary points-based empirical wavelet transform (DBPBEWT). The suggested approach is endorsed by using the publicly available dataset MIT-BIH Arrhythmia dataset. The simulation outcomes clearly demonstrates that the proposed technique outperforms other cutting edge techniques when compared via performance metrics namely output signal-to-noise ratio (SNRo) as well as correlation coefficient (CCE) at different intensities of input signal-to-noise ratio (SNRi).