DEEP LEARNING MODEL FOR CUFFLESS ESTIMATION OF BLOOD PRESSURE USING PPG
Abstract
Blood pressure is a vital sign used to measure the health of human heart. A persistent abnormal blood pressure can cause various cardiovascular diseases and if left untreated can lead to organ damage. Continuous and periodic blood pressure assessment is relevant for heart disease prevention. Regardless of the BP monitoring techniques in literature which are intermittent and cumbersome, several studies have considered using photoplethysmogram (PPG) signal as a cuffless and continuous measure of blood pressure. Here we propose a method to measure systolic BP using PPG and bidirectional GRU (Gated Recurrent Unit). The PPG signals of 219 subjects from the PPG-BP database are pre-processed using a zero-phase IIR Butterworth low pass filter. The pre-processed and normalized signal is then directly fed to the deep neural network architecture. The proposed work uses PPG signal with only 6 features along with Bi-GRU to estimate systolic BP. Certain PPG attributes such as cardiac period, systolic time, diastolic time, pulse area, systolic width and diastolic width are extracted from each cycle. The performance is measured based on metrics like mean absolute error (MAE) and standard deviation (SD). MAE of 4.56mmHg and SD of ±6.48mmHg are obtained for SBP meeting the requirement set by the AAMI standard.

Authors
S.J. Alphonsa Salu, D. Jeraldin Auxillia
St. Xavier’s Catholic College of Engineering, India

Keywords
Photoplethysmography, Systolic BP, Pulse Area, Cardiac Period, Bidirectional GRU
Yearly Full Views
JanuaryFebruaryMarchAprilMayJuneJulyAugustSeptemberOctoberNovemberDecember
000000000000
Published By :
ICTACT
Published In :
ICTACT Journal on Microelectronics
( Volume: 10 , Issue: 3 , Pages: 1839 - 1845 )
Date of Publication :
October 2024
Page Views :
12
Full Text Views :

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.