Detecting subclinical keratoconus
Keratoconus is a disease of the cornea, the transparent window of the eye, which primarily affects under 20 years olds. It’s relatively common in certain ethnicities and if left untreated can cause irreparable damage and lead to blindness. An efficient treatment now exists but it relies on early detection of the disease. However current automatic approaches do not use the latest imaging technology, anterior segment OCT, instead relying on derived parameters.
Our approach is to develop a CNN to distinguish subclinical (i.e early forms of the disease) from normals. We have access to large datasets of normal and subclinical keratoconus cornea imaging from Moorfields Eye Hospital (N>3000).