Diabetic Retinopathy affects over 94 million people worldwide. Diagnosis is severely constrained by a lack of specialist expertise, limited equipment access, and high screening costs. Early detection is critical to preventing vision loss, yet most cases go undiagnosed.
Developed a non-invasive automated screening system using color fundus photographs as input — eliminating the need for specialist equipment at point of capture.
Built an image analysis and pattern recognition pipeline to process retinal photographs and identify Diabetic Retinopathy markers with high accuracy.
Applied ML algorithms to assign probability scores indicating whether DR is present, enabling clinical triage and prioritization.
Designed for large-scale deployment, making automated screening accessible in underserved regions with limited access to ophthalmology specialists.
Enabled early detection of Diabetic Retinopathy, leading to proactive treatment and reduced risk of vision loss.
Highly effective at scale due to the automated, non-invasive methodology — no specialist equipment required at point of capture.
Democratized access to screening in underserved and under-resourced healthcare settings.