Contactless AI Support for Early Reading Difficulties

According to statistical estimates, dyslexia affects as many as 95,000 children in Serbia. For many of them, reading difficulties can make everyday learning more demanding and may also affect confidence, motivation, and social and emotional development. Since these difficulties can appear differently in each child, early recognition and support adapted to individual needs are especially important.

ReadTREAT — Computer-aided contactless assessment of reading difficulties and personalized treatment — develops objective, technology-supported methods for the early identification of reading difficulties and the personalized monitoring of treatment effects. The project included an entire generation of children from an entire second grade of a primary school, who were followed as they progressed into third grade. This enabled the team to observe reading development, individual needs, and the effects of personalized support over time.

The project combines reading tasks with non-invasive measurements to better understand how children read and how different forms of support may help them. Particular attention is given to individualized approaches, including adaptations in the way text is presented, as this can make reading easier and more accessible for some children. Through initial, personalized, interactive, and final screening phases, ReadTREAT supports the development of a tool for assessing reading difficulties, recommending individualized approaches, and tracking treatment outcomes.

Partners: School of Electrical Engineering – University of Belgrade, ICEF, Faculty of Philosophy – University of Belgrade, Singidunum University, and University of Hawaii.

This project is financed within the Diaspora Program of the Science Fund of the Republic of Serbia, #17473, supported by the SAIGE project of the World Bank.