EEG denoising using wavelet packet decomposition and independent component analysis

Abstract

EEG is a common-used instrument with plenty of biomedical and neuroengineering implementation. However the artifact contamination limits the practical use of EEG affecting feature extraction and classification. Various sophisti-cated methods are used to cleanse EEG, each offering a trade between computation speed and noise reduction ability. This work proposes fast denoising method based on a mixture of wavelet packet decomposition and independent component analysis.