Noise Removal in EEG Signals Using SWT–ICA Combinational Approach

Electroencephalogram (EEG) represents the electrical activity of the brain recorded by placing several electrodes on the scalp. EEG signals are complex in nature and consist of various artifacts like ocular, muscular, cardiac, etc. The artifacts removal in EEG signals can be majorly modeled by considering it of type Additive White Gaussian Noise (AWGN) in nature. Independent Component Analysis (ICA) is known for its ability to filter out the artifacts from the signal, and hence it is used to rearrange the source signal into two mixtures in a way that the brain signals and the artifacts get separated, although there is a constraint that ICA can only be performed on multichannel signal input. In the present case as the input EEG is single channel, hence, ICA is applied in combination with Stationary Wavelet Transform (SWT) for noise filtering of EEG signals. The quantitative evaluation of proposed approach has been made using Signal-to-Noise Ratio (SNR) parameter which depicts satisfactory filtering at varying intensity levels of AWGN.