EXTRACTION OF RESPIRATORY RATE FROM PPG SIGNALS USING PCA AND EMD

Abstract

Photoplethysmography is a non-invasive electro-optic method developed by Hertzman, which provides information on the blood volume flowing at a particular test site on the body close to the skin. PPG waveform contains two components; one, attributable to the pulsatile component in the vessels, i.e. the arterial pulse, which is caused by the heartbeat, and gives a rapidly alternating signal (AC component). The second one is due to the blood volume and its change in the skin which gives a steady signal that changes very slowly (DC component). PPG signal consists of not only the heart-beat information but also a respiratory signal. Estimation of respiration rates from Photoplethysmographic (PPG) signals would be an alternative approach for obtaining respiration related information.. There have been several efforts on PPG Derived Respiration (PDR), these methods are based on different signal processing techniques like filtering, wavelets and other statistical methods, which work by extraction of respiratory trend embedded into various physiological signals. PCA identifies patterns in data, and expresses the data in such a way as to highlight their similarities and differences. Since patterns in data can be hard to find in data of high dimension, where the luxury of graphical representation is not available, PCA is a powerful tool for analyzing such data. Due to external stimuli, biomedical signals are in general non-linear and non-stationary. Empirical Mode Decomposition is ideally suited to extract essential components which are characteristic of the underlying biological or physiological processes. The basis functions, called Intrinsic Mode Functions (IMFs) represent a complete set of locally orthogonal basis functions whose amplitude and frequency may vary over time. The contribution reviews the technique of EMD and related algorithms and discusses illustrative applications. Test results on PPG signals of the well known MIMIC database from Physiobank archive reveal that the proposed EMD method has efficiently extracted respiratory information from PPG signals. The evaluated similarity parameters in both time and frequency frequency domains for original and estimated respiratory rates have shown the superiority of the method.