French neonatologist Patrick Pladys leads the European research project DigiNew-B, which aims to detect infections in premature newborns. It uses both deep learning and artificial intelligence techniques.
He explained to Euronews how the device could save babies' lives.
"Why artificial intelligence? Because we're dealing with a complex problem. The clinician is usually able to respond to 2 or 3 signs that are abnormal and say "this baby is infected". With a premature baby, it is more complicated because the signs are discrete, diverse, there are many. And what we observe may be due to something other than the infection. To solve this problem we need the help of artificial intelligence because it allows us to analyse complex problems, that is to say, to integrate a whole set of parameters in the decision-making process, to be able to help the clinician in this decision-making process".
"First of all, we try to be non-invasive, not to intervene on the child but to use the data that is available. They are those of the monitors: heart and respiratory rate, oxygenation... We also have clinical data that we will capture on video. And then we have observation data from the nurse. We put all that data together and integrate a lot of babies in our study so that we can identify the babies who are at risk of infection and those who are not at risk. Based on that, if we project a child on these tests, over time he will have a probability of not being infected and then of being infected, and at a certain level the system will give the clinician an alert so he will be able to decide whether or not to start treatment".
"The ongoing evolution is mixing both technological progress and human progress, and the two are not mutually exclusive, on the contrary, I think that thanks to diagnostic assistance systems, we free up time for the human and for the relationship. Medicine is constantly advancing in both these aspects. There are technological advances and at the same time in neonatal, in the investment of parents - everything concerning the environment of the newborn on which we make a lot of progress. It takes both for medicine to progress".
This report was filmed before the coronavirus outbreak.