Detection of Retinal & Eye Diseases by using Convolutional Neural Network


Suresh Limkar, Arbaaz Shaikh, Kirti Chiplunkar, Meghana Kharache, Tejaswini Chaudhari, Sneha kulkarni
Department of Computer Engineering, AISSMS IOIT, Pune, India.


In todays generation, issues related to retinal diseases and eye diseases are increas- ing due to digitized world. Artificial Intelligence has provided a platform through which the early detection of diseases is possible and on the basis of that proper treatment can be made available. CNV (Choroidal neovascularization), DME (Diabetic macular edema), Drusen (reti- nal diseases), Cataract and Glaucoma (eye diseases) are the diseases considered for the pro- posed system. Cataract and Glaucoma are very common eye diseases leading to blindness. Cataract can only be diagnosed with Torch light examination, slit lamp examination etc. those required for Anterior Segment examination. Above diseases must be diagnosed at early stage to prevent complications. OCT (Optical coherence tomography) is an imaging technique which provides micrometer resolution images of retina The objective of this system is to improve the accuracy of the diseases detection like Cataract and Glaucoma along with retinal diseases CNV (Choroidal neovascularization), DME (Diabetic macular edema) and Drusen using the similar model as used for OCT images. Along with OCT images for detection of retinal diseases, eye scans are used for detection of Cataract and Glaucoma. Dataset of around 86,000 OCT, Cata- ract, Glaucoma and normal images was used. OCT, Cataract, Glaucoma and normal images were pre-processed and around 95% accuracy was achieved.