Abstract—Photophethysmography (PPG) is a technique used to measure and record changes in the blood volume of a part of the body and has become an essential tool in the practice of medicine because it is a non-invasive measure of low cost and easy implementation. The PPG signal has been widely used in clinical settings—emergency room, operating room, etc.—but reliable and accurate measurements are achieved only when the patient is at rest. Different indices for assessing the cardiovascular system can be obtained from the analysis of PPG signals. These indices are derived from the analysis of the shape of the PPG signals of patients at rest. When the patient moves, the shape and the spectrum of the PPG signal are altered. This paper presents a new method for the estimation of the main frequency component of PPG signals corrupted with motion artifacts. In addition, the analysis of the singular values obtained with this method allows to estimate the level of noise affecting the PPG signal. The periodicities of the noisy PPG signal can be estimated, simply by reconstructing the signal with its main frequency component and then, detect its zero crossings. The proposed method has a low computational cost. This article compares the performance of the method Singular Value Decomposition of the Time–Frequency Distribution (SVDTFD) with Discrete Wavelet Transform (DWT) and Ensemble Empirical Mode Decomposition (EEMD) by estimating the heart rate from PPG signals with motion artifacts produced by the continuous movement of patients.
 

Index Terms—photoplethysmography, motion artifacts, singular value decomposition, short–time Fourier transform, discrete wavelet transform, ensemble empirical mode decomposition

Cite: Juan F. Rojano and Claudia V. Isaza, "Singular Value Decomposition of the Time-Frequency Distribution of PPG Signals for Motion Artifact Reduction," International Journal of Signal Processing Systems, Vol. 4, No. 6, pp. 475-482, December 2016. doi: 10.18178/ijsps.4.6.475-482