Artifact removal of EEG data using wavelet total variation denoising and independent component analysis

Abstract

The Electroencephalogram (EEG) signals have very small amplitudes, which allow for the data to be readily contaminated by numerous artifacts. When it comes to clinical assessment, the presence of artifacts makes the study of EEG more complex. Power Line noise, eye movements, Electromyogram (EMG), and Electrocardiogram (ECG) are the most often seen artifacts that impact the EEG. Various researchers have developed a variety of strategies and procedures to deal with these artifacts. We provide a method for denoising the EEG signal in this work. The suggested method is implemented using a combined approach of wavelet total variation denoising method (WATV) and Independent Component Analysis (ICA). ICA technique entails running ICA algorithm on independent components to derive the components. In the case of artifactual events, just the wavelet-ICA components related to that event are used and then eliminated. To create artifact-free EEG, the artifact-free wavelet components are reconstructed. The complete approach may be confirmed for simulated signals and may be utilized for processing biological data, which may include EEG signal measurements, and for images, such as MRIs, contaminated by additional random noise. Signal to Noise Ratio (SNR) and Root Mean Square Error (RMSE) will be used to evaluate the algorithm’s performance. The WATV-ICA framework improves SNR more than the other techniques, according to simulation results.