Researchers at Moorfields Eye Hospital and UCL Institute of Ophthalmology have developed an artificial intelligence (AI) system called RETFound that can identify sight-threatening eye diseases and predict general health conditions like heart attacks, stroke, and Parkinson’s disease. RETFound is one of the first AI models in healthcare and the first in ophthalmology. It was trained using millions of eye scans from the NHS and is now being made available for free to institutions worldwide to help detect and treat blindness using AI. The system outperforms existing AI models and can work well in diverse populations and patients with rare diseases.

Senior author Professor Pearse Keane believes that if the UK combines high-quality clinical data from the NHS with top computer science expertise from its universities, it can become a world leader in AI-enabled healthcare. RETFound has the potential to be developed for hundreds of other sight-threatening eye diseases. The UK government views AI foundation models as transformative technology, and RETFound follows the launch of other foundation models like ChatGPT, which is a versatile language tool.

RETFound uses an innovative self-supervising approach to reduce the need for expert human labels, making it more efficient. It could significantly improve the diagnosis of eye diseases and predict systemic diseases like Parkinson’s, stroke, and heart failure. The eye is considered a window into overall health, and understanding the eye-body relationship is crucial in approaching complex diseases and the problems associated with aging.

RETFound has shown equal effectiveness in detecting diseases across diverse populations and has been used by researchers worldwide. It was trained with a dataset from Moorfields Eye Hospital and is available for use on GitHub. The project was a collaboration between various research centers and brought together teams from UCL and Moorfields.

>Source link>

>>Join our Facebook Group be part of community. <<

By hassani

Leave a Reply

Your email address will not be published. Required fields are marked *