Multi-level Motion Artifacts Reduction in Photoplethysmography Signal Using Singular Value Decomposition

Abstract

Photoplethysmography (PPG) is used for measuring vital cardiopulmonary indices such as heart rate and blood oxygen saturation (SpO2). But PPG signals get inevitably corrupted by movements of the patient and results in inaccurate calculation of heart rate and SpO2. In this paper, we report a method that uses a multi-level singular value decomposition (SVD) technique for effective reduction of motion artifacts while preserving the PPG morphology along with baseline. Results show impressive improvement on the signal quality and suppression of motion artifacts of the PPG signal. The PPG signals without motion artifacts obtained using out proposed method shows an average error of 0.69% in heart rate measurement with respect to the reference signal and an average max difference of 1.73% in the SpO2 estimation (in comparison to the average max difference of 1.69% for the reference signal). The proposed method has potential for accurate estimation of heart rate and SpO2 from PPG signal.