MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202641082776 A) filed by Parkkavi E; Nithyasri S K; and Ananthi S on July 06, 2026, for Retino Guard Ai: An Efficientnet Model And Deep Learning Frame Work Based Diabetic Retinopathy Detection System.
Inventors include Parkkavi E; Nithyasri S K; and Ananthi S.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: RetinoGuard Al is an integrated artificial intelligence system for automated diagnosis of diabetic retinopathy (DR) from retinal fundus images. The invention utilizes a pre-trained and fine-tuned convolutional neural network EffieientNet-B4 for DR classification with confidence score. A Grad- CAM visual explanation module provides saliency maps highlighting discriminative retinal regions. Furthermore, the invention provides a retinal region activation analysis module, which quantifies model attention on five anatomical retinal zones: optic disc, macula, superior arcade, inferior arcade, and peripheral retina. In addition, RetinoGuard AT features session-based user authentication module, which verifies the credentials of the requesting user and grants access only to registered users based on their roles. Finally, the invention provides a web application implemented as a Docker container that allows for convenient clinical deployment.
Disclaimer: Curated by HT Syndication.
