Hybrid Deep Learning Network to Classify Eye Diseases

Early identification and ongoing monitoring of
individuals with eye illnesses have been key concerns in
computer- assisted detection. In computer aid screens, detecting
one or more specific types of eye disease has achieved a big
advance. Glaucoma is a collection of eye illnesses that will
damage the optic nerve, which is sole responsible for information
sharing between eye to the brain, and can lead to blindness if left
untreated. Glaucoma must be detected as early as possible for
proper treatment. Diabetic retinopathy is also an eye illness
due to diabetes which in turn harm the blood vessels of retina.
This system proposed a Hybrid Ensemble Deep Learning system
for early detection of Glaucoma and diabetic retinopathy. Eye
images are first preprocessed to make them ready for subsequent
processing. Based on the features retrieved during training, the
system classifies incoming input images as normal eye, glaucoma
afflicted eye, or diabetic retinopathy affected eye. This system
produces 95% of accurate and stable output when compared to
existing system.
Index Terms—Deep learning, Hybrid Deep learning, Convolutional
Neural Networks, Glaucoma, Diabetic Retinopathy.