الفهرس 
Title : Contactless Measurement System for Vital Signs from Short-Time Facial Videos Using Signals Processing, Regression and Classification Models
SummaryOnly 14 pages are availabe for public view
 |  | from | 88 |  |  |
| | | from | 88 | | |
Abstract
Vital signs measurement is the basic step for any health check. They include heart rate,
respiratory rate, temperature, blood pressure and blood oxygen saturation. We proposed
measurement system using short time facial video and hybrid camera (RGB and
thermal camera). Temperature is measured using thermal camera, while other vital
signs only use RGB camera with standard camera specs. Methodologies used are facial
landmarks detection, image processing, signal processing, classification, and regression
models. System is evaluated against patient monitor readings using three limb
electrodes, blood pressure cuff, and pulse oximeter accessories. System is tested on 225
subjects with varying demographics. Heart rate is estimated with accuracy ± 6.16 beats
per minute, respiratory rate with ± 3.38 respirations per minute, SPO2 with ± 0.91%.
and blood pressure category with classification accuracy of 71%. Reference and
measured data statistics are shown and their correlations are investigated.