IIT Madras develops 1st India-specific AI model to determine foetus age

Researchers at IIT Madras have developed an AI model, ‘Garbhini-GA2’, specifically for the Indian population to accurately determine the foetal age in pregnant women, reducing errors in gestational age estimation by almost three times. The model has the potential to improve maternal and infant health outcomes in India.

The team, including researchers from THSTI, published their study in Lancet Regional Health Southeast Asia, highlighting the importance of accurate gestational age estimation for proper care of pregnant women and to determine delivery dates. The model is a part of the GARBH-Ini programme, aimed at enhancing public health in India using advanced data science and AI/ML techniques.

Garbhini-GA2, developed using GARBH-Ini cohort data from Gurugram Civil Hospital, Haryana, uses ultrasound parameters to accurately estimate foetal age in Indian population. Once validated in pan-India cohorts, the model can be deployed in clinics across the country to reduce maternal and infant mortality rates. Dr. Rajesh Gokhale from DBT India Initiative praised the development of population-specific models for gestational age estimation.