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{190754,
        author = {YASHWANTH PATEL GJ and IRFAN NABILAL RON and MANOJ C R and YASHASWIN M and Shaziya Banu},
        title = {Point-of-Care Diabetic Retinopathy Screening Using Edge-Optimized Deep Learning},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {8},
        pages = {4914-4920},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=190754},
        abstract = {Diabetic Retinopathy (DR) remains one of the leading causes of avoidable vision loss among individuals with diabetes, making early diagnosis and precise severity grading essential for effective clinical intervention. Conventional DR screening relies heavily on manual assessment of retinal fundus images by experienced ophthalmologists, resulting in high operational costs, limited scalability, and delayed diagnosis, particularly in resource-constrained healthcare settings. To address these challenges, this paper presents an automated Diabetic Retinopathy screening system based on a hybrid deep learning framework that combines DenseNet201 and ResNet50 convolutional neural networks. Retinal fundus images obtained from the Kaggle Diabetic Retinopathy dataset are utilized and categorized into five clinical stages: No DR, Mild, Moderate, Severe, and Proliferative DR. Transfer learning is employed by initializing both networks with pretrained ImageNet weights, while selectively freezing early layers to preserve generalized feature representations. High-level feature maps extracted from both architectures are fused through feature concatenation and processed by fully connected layers with a softmax classifier to enable multi-class severity prediction. The trained model is deployed using a FastAPI-based backend integrated with a web-based interface, facilitating real-time retinal image analysis. Experimental outcomes demonstrate that the proposed system offers an efficient, accurate, and practical solution for automated DR screening, supporting timely diagnosis and enhancing accessibility in clinical and point-of-care environments.},
        keywords = {Diabetic Retinopathy, Edge AI, Deep Learning, DenseNet, ResNet, Medical Image Analysis, CNN},
        month = {January},
        }