RetinaSight: A Clinician-in-the-Loop Platform for Robust, Explainable and Uncertainty-Aware Diabetic Retinopathy Screening and Risk Assessment

Vedhavarshini Kannan; P.Divyadharshini Palanikumar; P.M.G.Jegathambal Ganeshan

RetinaSight: A Clinician-in-the-Loop Platform for Robust, Explainable and Uncertainty-Aware Diabetic Retinopathy Screening and Risk Assessment

Author(s)	Ms Vedhavarshini Kannan, Ms. P.Divyadharshini Palanikumar, Ms. P.M.G.Jegathambal Ganeshan
Country	India
Abstract	Diabetic retinopathy is a major cause of preventable vision impairment, making timely retinal screening important for people with diabetes. This project presents RetinaSight, a web-based research prototype for organized, explainable diabetic retinopathy screening and future multimodal risk assessment. The application uses React and TypeScript for patient registration, examination management and visualization, while a Python FastAPI backend manages image uploads, SQLite records and PDF reports. A local dataset-import pipeline supports retinal photographs and five-level grading labels from sources such as APTOS 2019 and EyePACS, with separate pathways for external datasets. PyTorch scripts are provided for training candidate convolutional networks, including ResNet18 and EfficientNet-B0, and for generating checkpoint-dependent predictions, entropy scores and Grad-CAM visualizations. The prototype also supports technical image prechecks, clinician-review flags, longitudinal record comparison and quantitative evaluation from labeled prediction files. Fourteen backend tests passed during functional verification. However, no retinal dataset images or trained clinical model are bundled, and diagnostic accuracy, multimodal fusion, image gradability and external clinical validity remain unestablished. The system is intended for research and decision-support development, not autonomous diagnosis.
Keywords	Grad-CAM, FastAPI, EyePACS, APTOS, ResNet18, EfficientNet-B0
Field	Engineering
Published In	Volume 8, Issue 5, September-October 2026
Published On	2026-10-24

View / Download PDF File

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

RetinaSight: A Clinician-in-the-Loop Platform for Robust, Explainable and Uncertainty-Aware Diabetic Retinopathy Screening and Risk Assessment

Share this