Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{193259,
        author = {Cyrus Gomes and Bhakti Palghadmal and Gayatri Nadar and Sachin Narkhede},
        title = {ClearSight: Deep Learning-Based Classification of Retinal Diseases},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {9},
        pages = {4475-4481},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=193259},
        abstract = {Retinal diseases such as Diabetic Retinopathy (DR), Glaucoma, Cataract and Age-Related Macular Degeneration (ARMD) have been factors behind partial vision loss or even in serious cases complete blindness in our country. Often they are not caught at their infancy stages due to lack of testing, or simple oversight by the patients and doctors themselves. Early diagnoses and treatment can rectify and heal the eye before causing too much of a signifcant damage. However manual diagnoses can be tough as India lacks ophthalmologist per population head. We offer to solve this by creating a tool for diagnosing and reducing the workload on already encumbered health professionals. ClearSIght: a system built using deep learning, using multiple classes for classification of retinal diseases using the fundus images of the eye. Using preprocessing and training three models: ResNet, EfficientNet, InceptionV3 on robust, labeled dataset by healthcare professionals. Combined with ensemble to give accurate resultsm taking the mean proabbaility for the classes. We focus on three classes in this paper i.e Normal, Diabetic Retinopathy, Cataracts, Glaucoma, Macular Degeneration The system also incorporates understandable AI methods - particularly Grad-CAM - to offer clear visual explanations that emphasize the retinal areas affecting model predictions. A web interface built on Streamlit enables users to load fundus images and obtain instant disease identification accompanied by visual heatmaps and assurance scores. Tests show that the combined model reaches higher precision and clarity than single-model standard comparisons. Hence, ClearSight provides a dependable, straightforward and easily reachable AI-supported tool for screening retinal diseases - especially useful in areas with scarce ophthalmic knowledge.},
        keywords = {Exoplanets, Habitability, NASA Dataset, Habitable Zone, Atmospheric Composition, Gravitational Effects, Planetary Dynamics, Space Exploration, Three.js, Extraterrestrial Life},
        month = {February},
        }