Non-contact sleep monitoring and sleep apnea detection
Keywords: Sleep monitoring, ambient health monitoring, sleep apnea, computer vision.
Overview of Research
Obstructive sleep apnea (OSA) is a condition marked by reductions or cessations in breathing during sleep. These respiratory disturbances negatively impact sleep quality, causing daytime sleepiness and increasing the risk of car or occupational accidents. It is estimated to affect up to 8% of the adult population and has been linked to heart disease, obesity, depression and diabetes. Unfortunately, many individuals are unaware that they have OSA, with diagnosis rates lower than 20%. The current gold standard for OSA diagnosis is polysomnography (PSG), a comprehensive overnight recording of physiological signals involving the attachment of many sensors to the body. In addition to the discomfort and inconvenience associated with going to a sleep clinic for assessment, average wait times for PSG in Canada (with the exception of Ontario) vary between 8-30 months. The average cost of PSG is approximately $500. The restrictions associated with conventional OSA assessment highlight the need for a cost-effective, readily available alternative for sleep monitoring.
We are currently developing a non-contact system for sleep monitoring using a computer vision based approach. The system consists of an IR-sensitive video camera and a source of IR illumination. As the system is wireless and compact, it is convenient for home usage and has minimal discomfort for the user. The camera is mounted directly above the bed and captures the motion of the chest. We find the chest motion from the video and use it to determine respiratory rate and heart rate. Video is used only to extract chest motion – it is discarded immediately afterwards. The next step is to identify occurrences of OSA and test the clinical utility of the system for predicting OSA severity.
Videos
Noncontact Vision-Based Cardiopulmonary Monitoring in Different Sleeping Positions
Related Publications
M. Li, A. Yadollahi, and B. Taati. Non-Contact Vision-Based Cardiopulmonary Monitoring in Different Sleeping Positions. IEEE Journal of Biomedical and Health Informatics (IEEE JBHI), Vol. 21, No. 5, Sep 2017.
M.H. Li, A. Yadollahi, B. Taati. (2014). "A Non-Contact Vision-Based System for Respiratory Rate Estimation," 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), August 26-30, Chicago, IL, USA.
Research Team
Azadeh Yadollahi, Toronto Rehabilitation Institute
Babak Taati, Toronto Rehabilitation Institute
Cathy Zhu, Toronto Rehabilitation Institute
Maziar Hafezi, University of Toronto
Michael Li, University of Toronto
Nasim Montazeri, Toronto Rehabilitation Institute
Sina Akbarian, University of Toronto
