Contactless Patient Care Using Hospital IoT: CCTV-Camera-Based Physiological Monitoring in ICU

The continuous vital signs monitoring allows clinicians to timely assess the physiological conditions of the patient in intensive care unit (ICU). Internet of Things (IoT) in the hospital that integrated various sensors (e.g., cameras) may enable intelligent healthcare. In this article, we proposed a remote patient monitoring system that exploits closed-circuit television (CCTV) cameras in an IoT infrastructure as optical sensors for noncontact physiological measurement, extending its scope from surveillance to warding. The proposed monitoring system implemented the latest camera photoplethysmography algorithms for cardio-respiratory measurement, particularly focused on heart rate (HR) and breathing rate (BR) as fundamental biomarkers for earlier warning of deterioration. A clinical trial involving 25 critically ill patients was carried out in ICU, where the camera focal length (i.e., key impact factor) is thoroughly investigated. The clinical results show that our system achieves a mean absolute error (MAE) of 1.3 bpm for HR and 0.7 brpm for BR in the far-focus mode; an MAE of 1.0 bpm for HR and 1.8 brpm for BR in the near-focus mode, which are in the range of clinical acceptance. The comparison between the two modes suggests that the far-focus is more suitable for monitoring patient's vital signs in this scenario. The success rate of HR and BR in the far-focus mode is 94.5% (MAE = 2 bpm) and 96.7% (MAE = 3 brpm), respectively. The prototypes show that the increased measurement coverage and convenience of CCTV cameras in an IoT system are useful for ubiquitous patient monitoring in hospital care units.