Science Reporting

Voice recognition algorithm “translates” children’s crying

A group of researchers claim to have created a software tool based on artificial intelligence that is capable of distinguishing the cries of children. Very often parents find themselves unable to understand the real need for a crying child because they are not able to distinguish what can be considered as minimal nuances and changes in tone in the voice that, as several previous studies have shown, may indicate different needs.

The method is based on a modified language recognition algorithm so that it can work on crying. In addition to being useful in the home, it could also be useful in the health sector: for example, it could help doctors to understand the crying of children in the hospital or in serious conditions.

The study, published in the IEEE/CAA Journal of Automatica Sinica, explains how the algorithm is able to identify the patterns hidden in the cries of children, although the crying of each child can be considered as unique. There are in fact some basic characteristics that are common and that are dictated by the same needs or reasons.

The algorithm is independent, however, from the individual infant: it can be used universally since it can recognize and classify the various characteristics of crying and can also understand the level of urgency related to the need that the child intends to communicate.

“We are studying collaborations with hospitals and medical research centres, to obtain more data and input of scenarios and requirements, and we hope to have some products for clinical practice,” say the researchers in relation to the possibility of a practical use of this algorithm.