Diabetic Retinopathy Detection Using Image Analysis with Deep Learning

Correct diagnosis of retinal disease is important for timely intervention and effective treatment is extremely significant. Our approach is to improve image reconstruction using neural networks (CNN) and provide an accurate and practical method for eye disease detection. Retinal diseases such as glaucoma, diabetic retinopathy, high blood pressure, and atherosclerosis have many causes and require comprehensive evaluation. We focus on extracting key features of the eye, including eye structure, properties, and disease, using data from Research on the Retina (STARE) to train a CNN model that can identify a variety of eye diseases. Our approach includes extensive testing, using a variety of visualizations to illustrate the design, and integration into Django's website for deployment. The results were successful in classifying different types of retinal diseases, successfully detecting specific diseases, and demonstrating the potential of deep learning for modification of retinal diseases, thereby improving clinical outcomes and advancing the field of ophthalmology.