SINGAPORE EYE RESEARCH INSTITUTE In the medical field, however, he says that people have been“ a little slow,” to start using the technology – despite the fact that it can be used as a vital tool to combat what Jod describes as a“ tsunami of diabetes”.“ Worldwide, there’ s a whole tsunami of diabetes,” Jod says.“ And diabetic retinopathy is often one of the most severe complications of this as well.” The percentage of adults globally with diabetes stands at 14 %, says the World Health Organisation – a significant increase from 7 % in 1990. One of the most significant areas SERI has been working across with regard to AI imaging is diabetic retinopathy screening.“ One of the nice things about ophthalmology is that we are very heavy on imaging,” Jod explains.“ You can image the eye in unbelievable detail, more than in almost any other part of the body. The detail is unparalleled.”
Unlike expensive and cumbersome MRIs or CT scans, these highresolution ocular images are relatively cheap to take, providing the massive, high-quality data sets required to train sophisticated AI models. By implementing automated screening for diabetic retinopathy, AI performs a first pass – identifying early-stage disease in fundus photographs before a human doctor even sees them. This‘ first pass’ ensures that early signs of complications – which can be among the most severe results of systemic diabetes – are caught well before they become vision-threatening. By using algorithms to act as a virtual primary care physician, this system is able to automatically triage individuals to the correct specialised areas – helping patients receive the right treatment at the right time. seri. com. sg
9