Summary:
The pulse transit time (PTT) is commonly used to monitor pulse wave velocity (PWV). In general, the instruments of signal acquisition, from which these physiological variables are estimated, require a contact surface for the sensors' installation and positioning, such as an inflatable cuff, creating a restriction or obstruction to the users' movement and ergonomics. This paper describes the development and evaluation of a contactless cardiovascular monitor, which can measure the PTT and PWV by analyzing the photoplethysmographic (PPG) signal obtained by an RGB camera's green channel, i.e., without using sensors in contact with the skin. This monitor requires the PPG signal acquisition of two different regions of interest simultaneously: the forehead and right-hand palm. The time differences between two critical points of PPG signals were used to input machine learning algorithms, alongside other input features, such as user’s gender, height, and weight, to estimate Aortic PTT. The proposed monitor was tested by comparing its measurements of 36 healthy volunteers to the CARDIOS Dyna-MAPA+, gold standard equipment for these physiological variables measurement, showing a Pearson correlation coefficient of approximately 0.77 and a mean squared error of 0.104×10-3 for Aortic PTT.
Keywords: Blood Pressure, Camera, Machine Learning, Pulse Transit Time, Pulse Wave Velocity, Contactless.
DOI reference:
https://doi.org/10.14445/22490183/IJBTT-V11I1P602
Published on paper: March 2021.
Citation:
P.H. de Brito Souza, I. Machado Brito Souza, S. Gomes Soares Alcalá, P. Valverde de Oliveira Vitorino, A. Ferreira da Rocha, T.M. Barbosa, Video-based photoplethysmography and machine learning algorithms to achieve pulse wave velocity. International Journal of Biotech Trends and Technology. Vol. 11, nº. 1, pp. 7 - 15, March 2021.
Download this paper