A Camera-Based Pulse Transit Time Estimation Approach Towards Non-Intrusive Blood Pressure Monitoring

Abstract

Blood pressure (BP), as one of the primary vital signs, is of great importance to help assess an individual’s health conditions. Abnormal BP levels may provide clues to, or even induce, some life-threatening severe complications, such as heart attack, stroke, kidney disease, etc. Current BP measuring methods usually require the wearing of either an inflatable cuff or some contact sensors, which may limit the subject’s normal activities and greatly impact the user’s experience. We therefore aim to investigate a robust, non-contact, continuous and portable blood pressure and heart rate (HR) monitoring approach that features great usability and convenience for everyday use scenarios. To this end, in this paper, we propose a novel camera-based health monitoring system towards continuous BP and HR measurements based on pulse transit time (PTT) estimation and PPG extraction from facial regions. Our algorithm adapts the Laplacian pyramid based image processing technique to detect and capture the subtle changes caused by cardiac activities. To evaluate the usability and robustness of the proposed system, we conducted a pilot study with 4 subjects using the facial videos recorded from various mobile devices with different video recording positions. Specifically, compared with the reference readings acquired from a commercial blood pressure monitor, our system achieves the correlations in the range of 0.807 to 0.906 for HR estimation, and in the range of 0.64 to 0.94 for PTT estimates and systolic BP. It has been shown that the proposed system can provide accurate PTT estimations using regular mobile devices, which can be potentially used for non-intrusive continuous BP monitoring.