Eye Disease Prediction Using CNN

Abstract

In India, an estimated 15 million people are afflicted with blindness, and it is distressing to note that 70% of these instances could have been effectively cured at a precise point in time. The primary factor contributing to blindness is the absence of prompt medical intervention for initial-stage conditions. Accurate and early diagnosis is the only way to stop the progression of eye disorders. These disorders present themselves through a wide range of clearly observable symptoms. In order to obtain an accurate diagnosis, it is crucial to conduct a thorough investigation of these symptoms. This research supports the incorporation of digital image processing techniques, including segmentation and morphology, in addition to advanced deep learning approaches such as convolutional neural networks (CNN). Our proposed methodology presents a unique strategy for automating the detection of eye disorders by analyzing observable symptoms. The model specifically examines and classifies four common eye conditions: strabismus, exophthalmos, cataracts, uveitis, and conjunctivitis. The deployed deep neural network model is crucial in the timely identification of eye problems, offering a proactive approach to detect their presence. Moreover, the concept promotes individuals to actively pursue screening from ophthalmologists when it is deemed required. This fusion of image processing and deep learning technology offers as a promising approach to boost eye healthcare and reduce preventable blindness in India.