Contact-Free Physiological Monitoring of Cardiorespiratory States Using Radar and Optical Sensors

Abstract

The paper summarizes a Doctoral Thesis that focuses on two new approaches for unobtrusive contact-free monitoring of cardiorespiratory and hemodynamic states. First approach is based on radar signals and proposes a novel branched neural network architecture for classification of hemodynamic scenarios. The second is based on RGB camera signals and proposes multi-wavelength depth-dependant photoplethysmogram reconstruction, allowing for single-site pulse transit time measurement and blood pressure estimation using a consumer camera.