Dr. Siren Rettedal

Using AI to learn from births and resuscitations


Deprivation of oxygen to a newborn during and after birth can lead to birth asphyxia, a leading cause of newborn death, cerebral palsy, and other long-term injuries. Resuscitation activities include stimulation, clearing the airways, and face-mask-ventilation. In Norway, approximately 10% of term newborns need stimulation and around 3-4% need face-mask ventilation. Face-mask ventilation should begin within one minute after birth.

The Newborn Time project was developed to automatically create a timeline of birth and resuscitation activities. Artificial Intelligence (AI) is used to detect time of birth from thermal cameras mounted in the delivery room. Activity recognition is performed using AI in the form of deep convolutional neural networks (CNN) on thermal and RGB video from the resuscitation. The system is designed to recognize multiple time overlapping activities. The AI models are robust, reliable, general, and adaptive for use at different hospitals and settings.

The product is a collaboration between the University of Stavanger (UiS), Stavanger University Hospital (SUS), Laerdal, and BitYoga. UiS, SUS, and Laerdal have extensive experience in collaborative research on newborn care. Laerdal has developed Newborn Time data collection and technology which will be used at SUS. UiS developed site-adaptive AI methods for time of birth and resuscitation activities recognition from video. BitYoga delivers GDPR compliant consent technology for expecting mothers.