MUMBAI, India, July 13 -- Intellectual Property India has published a patent application (202641082575 A) filed by Srm University on July 04, 2026, for A System And A Method For Glaucoma Screening From Retinal Fundus Images.
Inventors include Chodavarapu, Tejaswi; Poola, Rahul Gowtham; Bethalam, Venkata Siva Sai Greeshma; Galimutti, Raju; Vaddi, Ramesh; and Yellampalli, Siva Sankar.
The application for the patent was published on July 10, 2026, under issue no. 28/2026.
Abstract: ABSTRACT A SYSTEM AND A METHOD FOR GLAUCOMA SCREENING FROM RETINAL FUNDUS IMAGES The present disclosure envisages a computer-implemented method (100) and system (200) for glaucoma screening from retinal fundus images. The method (100) includes receiving (102) unlabeled retinal fundus images and a machine-learning model trained using labeled image data. For each unlabeled image, a predictive output is generated (104), and a predictive uncertainty score is determined (106) using Shannon entropy. A Grad-CAM heatmap is generated (108) from an intermediate layer of the model to derive an explainability score. A forgetting score is determined (110) based on a difference between outputs of a student model and a teacher model. A composite acquisition score is computed (112) using the uncertainty score, explainability score, and forgetting score. The images are ranked and selected for annotation based on the composite acquisition score. Subsequently, the model is incrementally updated using the annotated images through a continual learning framework employing experience replay, elastic weight consolidation, and knowledge distillation.
Disclaimer: Curated by HT Syndication.
