Enhanced Empirical Mode DecompositionApproach toEliminateMotion Artifacts in EEG using ICA and DWT

To ease the concept of diagnosing human health a strong and viable biomedical signal is of importance. Biomedical signal measurement and processing of signal cause the probability of artifacts which can obstruct the features of interest and quality of information available in the signal. So elimination of artifacts from physiological signals is an essential step. The single channel measurement is important when instrumentation complexity is needed to be minimized, in spite of many multichannel signals recording methods available.In this paper, an enhanced empirical approach to remove the artifacts of single channel signal is described followed by filtering mechanism using ICA and DWT. The input EEG is a single channel and is converted into multichannel for ICA operations. The multi-channel EEG signal is filtered with fast ICA algorithmand DWT is employed to reject any traces of artifacts left in the signal. This technique is tested against currently available wavelet de-noising and EMD-ICA technique using functional near infrared spectroscopy (EEG) signal. Artifact removal technique has been evaluated by DSNR, Lambda (λ), Autocorrelation and PSD. The results pronounce the eligibility of proposed algorithm to stand on top of currently deployed algorithms with 12% improvement in DSNR and alsoa significant improvement in different parameters too.