Robust Non-Contact Monitoring of Respiratory Rate using a Depth Camera.

Area of Science:

• Biomedical Engineering
• Medical Devices
• Physiological Monitoring

Background:

• Respiratory rate (RR) is a critical vital sign for identifying acute illness and potential complications.
• Timely detection of RR changes facilitates prompt clinical intervention, improving patient outcomes.
• Current monitoring methods may require physical contact, posing limitations in certain clinical scenarios.

Purpose of the Study:

• To evaluate the performance of a depth-sensing camera system for continuous, non-contact monitoring of respiratory rate.
• To assess the accuracy and reliability of touchless RR measurement across various physiological conditions.

Main Methods:

• Healthy subjects breathed at set rates from 4 to 40 breaths/min.
• Depth data was captured using an Intel D415 RealSense camera.
• Real-time processing extracted torso depth changes to calculate respiratory rate (RR_depth).

Main Results:

• Achieved an overall Root Mean Square Deviation (RMSD) accuracy of 0.69 breaths/min.
• Bland-Altman analysis showed limits of agreement between -1.42 and 1.36 breaths/min.
• High accuracy (RMSD < 1 breath/min) was maintained across low, normal, and high respiratory rate sub-ranges.

Conclusions:

• The depth-sensing camera system demonstrates high accuracy for respiratory rate monitoring.
• The system effectively measures both high and low respiratory rates, clinically significant ranges.
• This touchless technology offers a promising approach for continuous physiological monitoring.