HHT based signal decomposition for reduction of motion artifacts in photoplethysmographic signals

Abstract

Motion artifact (MA) corrupted photoplethysmographic (PPG) signals are the main source of errors in the estimation of arterial blood oxygen saturation (SpO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> ) in pulse oximeters. For addressing the issue of MA reduction in pulse oximetry applications, the physical origins of PPG signals are to be explored and effective signal processing technique may be employed. In this paper, we propose simple and efficient empirical mode decomposition (EMD) method based on the Hilbert-Huang Transform (HHT) for MA reduction in PPG signals. EMD is relatively a new time-frequency analysis technique having wide range of applications. EMD uses HHT calculation to handle non-linear and non-stationary data to find the intrinsic mode function (IMF) components and analyze the variations in power spectrum over time. The efficacy of the proposed method is proved by comparing it with well known wavelet transform based MA reduction method for the PPG data recorded with different MA (Horizontal, Vertical and Bending motion of finger). While statistical analysis demonstrated the robustness of the method, the SpO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> estimations from MA reduced PPG signals by proposed method being very close to the actual ones, make it reliable for pulse oximetry applications.