Non-contact (touchless) monitoring of respiratory rate in a challenging anesthesia setting using a depth camera

We have developed a non-contact ("touchless") system based on depth-sensing camera technology for continuous monitoring of respiratory activity. Previous work from our group has demonstrated high accuracy of the system in monitoring a wide range of respiratory rates and signal morphologies across diverse conditions, including variations in lighting, posture, and coverings. Here, we report on the system's performance in a significantly more challenging anesthesia environment which included a wide range of respiratory rates and respiratory patterns, spontaneous and hand ventilated breathing, patient motion and caregiver interactions in the scene, and, in some cases, the presence of warming blankets covering the torso.

Data was collected opportunistically from 34 healthy volunteers from two separate studies, both of which had the primary objective of investigating the relationship between depth of anesthesia monitoring and anesthetic agents (inhaled and intravenous) across a wide range of anesthetic concentrations and hypnotic states. Depth-sensing information was acquired using an Intel D415 RealSense™ camera and processed to extract frame-by-frame depth changes within the subject's torso region corresponding to respiratory activity. A respiratory rate (RR_depth) was calculated and output once-per-second from the device. This was compared to a combined reference (RR_ref) derived from both a capnograph and an impedance-based respiratory monitor. Three time periods were evaluated: pre-anesthesia, intra-anesthesia and post-anesthesia.

The overall RMSD accuracy [bias] obtained for the combined data set was 1.92 [0.30] breaths/min. The performance results stratified according to pre-, intra-, and post-anesthesia stages were 1.71 [0.15], 1.95 [0.39] and 2.13 [0.08] breaths/min, respectively.

We have demonstrated the ability to continuously track respiratory rate during challenging conditions within an anesthesia setting using our non-contact, touchless, monitoring technology. We believe that our findings support the potential utility for continuous non-contact monitoring of respiration in clinical areas, such as the post-anesthesia care environment.