Project: #59
Non contact monitoring of neonatal health in ICU
The project aims to develop methods for reliable assessment of health and vital parameters using non contact methods in a neonatal ICU environment.
The research challenge is effective fusion of video streams from multiple visible light and IR cameras, and continuous identification of the optimal region of interest on the neonatal skin surface for reliable assessment of skin perfusion changes for HR estimation, chest motion patterns for respiratory health estimation, and breath patterns for RR estimation etc. Efficient data fusion is needed to improve signal reliability and ensure good spatio temporal coverage of the unobtrusively estimated vitals - a problem that is as yet uncracked. Moreover, the project will also explore AI based pose detetion and behavioral assessment methods to give an overall state of health of the baby - which goes beyond just the essential vital signs.