Comparison of PPG and BCG Features for Camera-based Blood Pressure Estimation by Ice Water Stimulation

Abstract

Non-invasive Blood Pressure (BP) measurement is highly demanded for pervasive healthcare with the development of Internet of Things, sensors and mobile technology. Camera-based Photoplethysmography (camera-PPG) has been applied for non-contact BP estimation. Most camera-PPG based approaches calculate the Pulse Transmission Time between different peripheral sites like face and palm for BP calibration, which require more than one body part to be simultaneously measured and thus introduce inconvenience to real applications. In this study, we investigate the feasibility of measuring BP from a single body site using either the forehead PPG signals or neck ballistocardiographic (BCG) motion signals. Two morphological features (K-value and Augmentation Index) that have clinical meanings for BP monitoring have been compared. The study was conducted in the ice water stimulation experiment involving 16 healthy subjects. The results show that the neck can be an attractive site for BP estimation as the neck-BCG signals show more distinct features (e.g. dicrotic wave) that have stronger correlations with BP than the forehead-PPG signals, and it eliminates the privacy issue of imaging a face. Both the K-value and Augmentation Index can well track the changes of BP. The conclusions drawn from this study inspire the selection of physiological site and features for non-contact BP estimation.