Removal of Narrowband Interference (PLI in ECG Signal) Using Ramanujan Periodic Transform (RPT)

Abstract: Suppression of interference from narrowband frequency signals play vital role in many signal processing and communication applications. A transform based method for suppression of narrow band interference in a biomedical signal is proposed. As a specific example Electrocardiogram (ECG) is considered for the analysis. ECG is one of the widely used biomedical signal. ECG signal is often contaminated with baseline wander noise, powerline interference (PLI) and artifacts (bioelectric signals), which complicates the processing of raw ECG signal. This work proposes an approach using Ramanujan periodic transform for reducing PLI and is tested on a subject data from MIT-BIH Arrhythmia database. A sum ($E$) of Euclidean error per block ($e_i$) is used as measure to quantify the suppression capability of RPT and notch filter based methods. The transformation is performed for different lengths ($N$), namely $36$, $72$, $108$, $144$, $180$. Every doubling of $N$-points results in $50{\%}$ reduction in error ($E$).