Automating sleep stage classification using Contactless BCG Sensor

It is estimated that 5.4 million Canadian adults have chronic sleep abnormalities. Symptoms are not visible to patients because they happen during the night. Hence, they remain undiagnosed. Besides, sleep abnormalities can cause different chronic health problems, that is sleep apnea, diabetes, stroke, brain injury, Parkinson’s disease, depression, and Alzheimer’s disease. Thus, measuring sleep behavior can diagnose sleep disorders and enable the early detection of other health conditions. Current sleep monitoring systems are expensive, labor-intensive, complex. Also, it is not possible to emulate the usual sleep environment in a sleep laboratory. Furthermore, manual scoring has considerable inter-scorer and intra-scorer variability, making its reliability and reproducibility questionable. That said, in this project, we propose a deep learning-based method for the automatic detection of sleep architecture using a contactless system that is based on the ballistocardiographic principle in an attempt to address one of today’s health care issues.

Faculty Supervisor:

Bessam Abdulrazak

Student:

Partner:

Mediterranean Institute of Technology

Discipline:

Computer science

Sector:

Health and Related Sciences & Technology; Artificial Intelligence; Information and Communications Technology

University:

Université de Sherbrooke

Program: