Skip to content
ZC
All projects
Deep learningComputer vision

Ocular Disease Intelligent Recognition (3rd prize winner)

In ophthalmology, fundus screening is an economic and effective way to prevent blindness as early as possible that cased by diabetes, glaucoma, cataract, age-related macular degeneration (AMD) and many other causes. The figure below shows five simulated eye conditions when they have corresponding disease compared with normal vision.

In ophthalmology, fundus screening is an economic and effective way to prevent blindness as early as possible that cased by diabetes, glaucoma, cataract, age-related macular degeneration (AMD) and many other causes. The figure below shows five simulated eye conditions when they have corresponding disease compared with normal vision.

Early detection of common ocular diseases is quite difficult since few symptoms are visible in the early stage of diseases. For example, people with longer standing diabetes have a higher chance of developing Diabetic Retinopathy (DR) and Diabetic Macular Edema (DME). The early sign of diabetic eye disease is microaneurysms, which is small and hard to detect. Macula is responsible for the central vision. Macula with fluid buildup will distort vision. Function degeneration of macula in aged people has no symptoms in the early stage. Cataract is also commonly seen in aged people. It will decrease human’s vision. Glaucoma refers to a group of diseases that causes damage to the optic nerve head, which is irreversible. Moreover, hypertension will change the morphological structures of blood vessels, for example, diameter changes and tortuosity alteration. It can produce cardiac-cerebral vascular diseases, like strokes and heart attacks. Finally, vision loss is also of high risk in myopia people because of progressive retinal pigment epithelial thinning and attenuation. Early detection of these diseases could prevent vision damage and other problems.

The purpose of this challenge is to develop methods for automatic ocular disease classification. With color fundus images from left and right eyes as inputs (may use other provided information, such as patient age, sex), the goal of ocular disease classification is to classify a patient into eight catagories, including normal, diabetes, glaucoma, cataract, AMD, hypertension, myopia and other diseases/abnormalities.