Creating and sharing knowledge for telecommunications

Assessment of Empirical Mode Decomposition Implementation in Cardiovascular Signals

Pinheiro, E.C. ; Postolache, O. ; Girão, P.M.

Assessment of Empirical Mode Decomposition Implementation in Cardiovascular Signals, Proc IMEKO TC4 Symp., Kosice, Slovakia, Vol. I, pp. xx - yy, September, 2010.

Digital Object Identifier:

Abstract
Biomedical signals are relentlessly superimposed with interferences and noise. The nonlinear processes which generate the signals, and the interferences, regularly exclude or limit the usage of classical linear techniques, and even wavelet transforms, to decompose the signal.
Empirical Mode Decomposition (EMD) is a recently proposed method for analyzing signals from a nonlinear viewpoint. EMD is not defined by a formal mathematical analysis, instead the decomposition is obtained by following an algorithm, requiring experimental investigation, thus being a fully data-driven technique.
Hence, this work evaluates the impact of EMD implementation in the processing of biomedical signals, namely on electrocardiogram, impedance and photoplethysmogram, and ballistocardiogram. All these signals are very sensible to motion artefacts, and therefore mode decomposition is important to separate the representative component from the pure noise. EMD was never applied to most of these signals, which are very physiologically meaningful, so the implementation’s impact is assessed.