Grouping in AI
acute care for the child

Video diagnosis of children suffering from respiratory failure

There is no clinically usable non-contact monitoring method for measuring the extent of respiratory distress in children, i.e. measurement of tidal volume (Vt), respiratory frequency (RF) and signs of respiratory retraction. Our research group is developing a solution to meet this need. The RETRACT system is a non-contact respiratory status monitoring system currently under development, based on 3D video combined with thermography.

The development of the algorithms is based on a clinical database currently comprising 200 patients, funding from the Canada Foundation for Innovation, which is financing the installation of these cameras in the pediatric intensive care unit at CHU Ste-Justine ($2.4 million), and the FRQS Chair in Artificial Intelligence, which is funding the students working on the project.

This project began in 2016 and is now at TRL 4 with a hardware component comprising a 3D kinect azure camera, a Lepton 3.5 infrared camera, a Jetson card and a calibrated support (synchronization of the cameras, see figure opposite) and an algorithmic component measuring tidal volume, respiratory rate and clinical signs of retraction.