Contactless extraction of respiratory rate from depth and thermal sensors
Metadata[+] Show full item record
Monitoring of respiration and restless sleep can help detect sleep disturbances that may be indicative of poor health and functional deficits. The current methods of estimating the respiratory rate such as Pneumograph, Capnograph, Photo-plethysmograph (PPG), Respiratory inductance plethysmography (RIP), involve sensors that are in contact with the patient. However, we have a few scenarios such as in hospitals and senior retirement communities where we would like to non-invasively collect the respiration rate and restless body motion where we are not able to place these types of sensors on patients. The initial requirement was to non-invasively monitor vital activity of patients in psychiatric centers. This work investigates a novel approach to estimate the respiratory rate of a person lying on the bed using depth and thermal sensors along with other signal processing algorithms. The initial proof of concept tests were conducted on three subjects. Additional testing on a diverse group of ten participants (ranging in age and body type) was performed to validate the algorithm and the data collection method. The depth and thermal waveforms captured were tested to explore a new approach for detecting individual respiratory rate noninvasively, using various algorithms to detect the region of the bed, common grids where a person is present, best signal selection from grids, and accurately estimate the respiratory rate and amount of body movement during sleep. The performance results at approximately 30 frames per second for the set of 10 participants was a mean error difference of 0.6 breaths per minute for the time domain algorithm and 0.8 breaths per minute for the frequency domain algorithm.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License.