Denoising of ECG Signal using Computationally Efficient Linear Phase Filtering

Abstract

An Electrocardiogram (ECG) signal is often contaminated with several types of noises, which may cause incorrect diagnosis of a heart patient. Therefore, the elimination of noise is always the first stage for the analysis of an ECG signal. In this work, high-frequency, power line interference, and baseline wander noise are eliminated using the frequency-response masking (FRM) technique and interpolated finite impulse response (IFIR) filter. These techniques are used because of computational efficiency and linear phase property. These techniques involve two or more finite impulse response (FIR) filters. The weighted least squares (WLS) regression method is used to design the sub-filters of IFIR and FRM filter. First, the computational filters are designed, and then these filters are used to eliminate the power line interference, high-frequency and baseline wander noise. © 2019 IEEE.