In 2018, Jeff received the first Fetal Age Machine Learning Initiative (FAMLI) grant, a feasibility investment concerned primarily with automated estimation of gestational age. It focused upon generation of specific ultrasound datasets to train machine learning models and on scoping additional functions and features of an AI-enabled ultrasound system.
In 2019, a second FAMLI grant extended the investment in data collection, expanded the resources available for machine learning, and introduced new partnerships with companies developing low-cost, handheld ultrasound solutions. As a result of this work, UNC developed a successful blind sweep gestational age estimation model that outperforms expert sonographers and has just launched a field validation of this gestational age tool in Chapel Hill and in Zambia.
The new FAMLI 3 award provides $17 million over 4 years for UNC to significantly expand this work. The primary goal of this award is the development and validation of one or more multi-functional, AI-enabled, automated, low-cost ultrasound devices, and the collection of data necessary to support widespread recommendation and adoption by agencies such as the World Health Organization and national Ministries of Health. This award will integrate with and leverage other Gates Foundation-funded work to expand the scope of the learning and testing datasets and support broad scientific engagement with international collaborators.
–Leslie H. Nelson-Bernier
Chief Philanthropy Officer, UNC Health
President, UNC Health Foundation
