MUMBAI, India, Oct. 12 -- Intellectual Property India has published a patent application (202631118425 A) filed by Nikhil Ranjan Nayak; M. Iqbal Jeelani; Dr. Mansha Gul; Dr Harpreet Singh; Dr. Manoj Patel; Dr Pasupuleti Sreenivasa Rao; Mrs. Swati Nitin Kakad; Dr Charanjeet Kaur; Girish Joshi; and Dr Pranay Wal on October 03, 2026, for A Deep Learning Method And System For Multi-Disease Detection From Retinal Fundus Images.

Inventors include Nikhil Ranjan Nayak; M. Iqbal Jeelani; Dr. Mansha Gul; Dr Harpreet Singh; Dr. Manoj Patel; Dr Pasupuleti Sreenivasa Rao; Mrs. Swati Nitin Kakad; Dr Charanjeet Kaur; Girish Joshi; and Dr Pranay Wal.

The application for the patent was published on October 09, 2026, under issue no. 41/2026.

Abstract: ABSTRACT The present invention relates to a deep learning method and system for multi-disease detection from retinal fundus images. The developed system receives a retinal image and initially performs automated image-quality assessment to identify images unsuitable for reliable analysis. The acceptable image is subjected to retinal-field detection, normalization, illumination correction, contrast enhancement and other preprocessing operations. A deep-learning network subsequently extracts global retinal features while an attention-based branch extracts localized disease-associated features. The global and localized representations are fused and supplied to a multi-label classification network configured to independently determine disease-specific probabilities. The system can simultaneously screen for diabetic retinopathy, hypertensive retinopathy, glaucoma, age-related macular degeneration, retinal vascular abnormalities and other selected retinal conditions. Class imbalance is addressed using weighted and/or focal learning strategies and disease-specific thresholds are optimized using validation data. An explainability module generates a visual heatmap identifying retinal regions contributing to each prediction. The invention provides a unified, scalable and clinically oriented computational framework for multi-disease retinal screening and can be deployed in ophthalmic clinics, hospitals, teleophthalmology platforms and resource-limited screening environments.

Disclaimer: Curated by HT Syndication.