Eliminate Artifact on ECG Recording Using the Soft Threshold Setting on Wavelet Coefficients at Independent Components of ICA

A common problem in ECG signal acquisition is the removal of artifacts and undesirable components to obtain a clean ECG signal, which helps to increase the accuracy of the clinical diagnosis process.However, with expecting received ECG in the high accuracy, the basic filters are not enough, because the ECG signal recording's often affected from differential sources with varying amplitudes and frequencies; furthermore, the recording process needs to be implemented via electrodes on the skin, which not only record the electrical activity of the heart, but also many other participating components such as Respiratory, Electroencephalogram (EEG), electrooculography (EOG), electromyography (EMG) with many artifact from outside.Therefore, conventional filters didn't meet the requirements of removing most of the impacting artifact components.In this study, the author has proposed a new method, that is to apply the independent component analysis (ICA) -combining wavelet transforms on each independent components to remove the abnormal noise, especially EMG to improve the accuracy of ECG signal recording with a correlation value of up to 0.971 compared to the desired.