Wavelet Decomposition Based Detection of Blood Pressure Level and Heart Rate Using Photoplethysmography Signals

Abstract

In this work, a new wavelet decomposition based algorithm has been proposed to predict systolic and diastolic blood pressures and heart rate changes. Instead of traditional measurements methods, Photoplethysmography (PPG) signals have been used for prediction. The easy, fast and non-invasive recording technique of PPG signals makes them promising for medical applications. PPG signals, blood pressure and heart rate information used in this work were obtained from an open database. The database contains recordings of PPG signals, blood pressure and heart rate information of 219 people which has been measured with traditional methods. We proposed to use Tunable Q-factor Wavelet Transform (TQWT) for signal decomposition. TQWT has been applied separately to each PPG signals and the average power, mean absolute deviation, skewness, kurtosis and standard deviation coefficients of each sub-band have been determined. Regression analysis has been performed by applying Artificial Neural Networks (ANN), Random Forest (RF) and Support Vector Machines (SVM) algorithms to the features of all sub-signals. Results of the regression analysis show that blood pressure and heart rate estimations with high correlation coefficients and low error rates have been obtained whit the RF algorithm.