Denoising of multichannel MCG data by the combination of EEMD and ICA and its effect on the pseudo current density maps. (April 2015)

• Record Type: Journal Article
• Authors: Mariyappa, N.; Sengottuvel, S.; Rajesh Patel, ; Parasakthi, C.; Gireesan, K.; Janawadkar, M.P.; Radhakrishnan, T.S.; Sundar, C.S.
• Is Part Of: Biomedical signal processing and control. Volume 18(2015)
• Page Start: 204
• Page End: 213
• Publication Date: 2015-04
• Subjects: Preprocessing -- Empirical mode decomposition -- Intrinsic mode functions -- Wavelet-independent component analysis -- Magnetic field map -- Pseudo current density map; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2014.12.012
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract : Highlights: The combination of EEMD and ICA proposed and used for multichannel MCG signal denoising. Reduces the computational complexity associated with EEMD alone and effectively enhances the signal quality. The effect of ICA alone, wICA, and EEMD-ICA on the magnetic field map and pseudo current density maps is investigated. The EEMD-ICA combination yield robust source estimate on the PCD map. The EEMD-ICA can be applied to beat-to-beat analysis of MCG signals. Abstract: The signal preprocessing is prerequisite for reduction of noise and for better estimation of sources from the measured field distribution of multichannel data, since different measurement channels may be contaminated by different types of artifacts and noise. Toward this, we use a combination of independent component analysis (ICA) and ensemble empirical mode decomposition (EEMD) to denoise the multichannel magnetocardiography (MCG) data. In this technique, MCG time series data is first subjected to ICA to obtain the statistically independent components (ICs) and subsequently the EEMD-interval threshold based denoising is applied to the ICs prior to the reconstruction of the signal. We compare the results obtained from EEMD-ICA with those obtained using the conventional ICA alone and also using the wavelet enhanced ICA (wICA). We illustrate the effect of these denoising techniques on the pseudo current density (PCD) maps, which aid in visualizing the source location. The results obtained from the EEMD-ICA are seen to be decidedly superior compared to those obtained by ICA alone and wICA methods.