A study on electrocardiogram signals denoising using s-transform & modified s-transform

Abstract

Electrocardiogram (ECG), a non-invasive technique is used as a primary diagnostic tool for cardiovascular diseases. It plays an important role in monitoring of patient and diagnosis attributable to its ease of use and non-invasive nature. Today, ECG signal processing is one of the challenging computational processes as different noises get embedded with it such as channel noise, electrode motion, muscle artifacts and baseline wander during the acquisition and transmission of signals. A cleaned ECG signal provides required information about the electrophysiology of the heart diseases and ischemic changes that may happen thereby giving valuable information about the functional details of the heart. The aim of this thesis is to remove the noise components from the ECG signal based on timefrequency domain representation using a denoising technique called S-transform. S-transform has a time-frequency resolution which is far from ideal. In this work, S-transform represents the noisy ECG signal in the time-frequency domain. Next, masking and thresholding technique is applied to remove the unwanted noise components from time-frequency domain. But, a noisy time series, with both signal and noise varying in frequency and in time, presents special challenges for improving the signal to noise ratio. This thesis proposes a modified S-transform, which offers better time frequency resolution compared to the original S-transform. This is achieved through the introduction of a new scaling parameter for the Gaussian window used in S-transform. This method gives good performance with high SNR value for different noises as compared to S-transform.