Filtering Frameworks for Removing Noise and Baseline Wander in ECG Signal

Abstract

Electrocardiogram (ECG)is the most basic biosignal tool in automated wireless patient monitoring system to diagnose and measure the electrical activity of the heart. Automated wireless patient monitoring system demands wireless and noise-free ECG signal transmission foraccurate diagnosis of cardiovascular diseases (CVDs). Noise is evidently present in wireless systems and eliminating noise from the original signal becomes essential for effective and accurate detection of ECG signal. Several techniques have been developed for effective filtering of ECG signal. In this article, we present different filtering frameworks for removing noise of ECG signal. Here, four filters, namely: Butterworth, Chebyshev, Elliptical and Savitzky-Golay have been implemented for noise extraction using MATLAB simulation software on ECG signals from various databases for their performance analysis. Further, detrends in the ECG signal generated due to baseline wander has been removed using a low order fit polynomial. Average SNR values of Savitzky-Golay filtered ECG signals shows best performance in noise removal as compared to other filters.