A SMART CRIB EMBEDDED WITH AI ALGORITHMS FOR AN EARLY DETECTION OF AUTISM SPECTRUM DISORDER IN PRETERM INFANTS
Project The SMARTER-ASD project aims to design, implement, and validate a smart crib for analyzing visual preference
and motor patterns in preterm infants in NICUs. The smart crib integrates a monitoring framework, consisting of
RGB-D cameras, sensorized pads and sensors already available in NICUs (e.g., cardio-monitor for physiological
data). The framework also integrates Electronic Health Records (EHRs), for retrieving infants’ co-morbidities and
clinical treatment. The smart crib is equipped with an Artificial Intelligence (AI) engine, for processing all this
data and providing healthcare professionals with quantitative means for early ASD screening in NICUs. This will
allow clinicians to perform infants’ active monitoring until an age when ASD diagnostic means may be considered reliable. Trustworthiness of AI algorithms will be prompted in all the algorithm lifecycle. To develop and test the smart crib, a study will be put in place relying on the collaboration with a major center specialized in preterm infants’ care and neonatal motor and visual behavior. To achieve its goal, SMARTER-ASD will bring together world-renowned healthcare professionals, computer-science researchers, biomedical engineers, and ethicists.