MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641048599 A) filed by Srm Trp Engineering College on April 16, 2026, for Energy Efficient Edge Based Compressed Vision Transformer System And Method For Real Time Diabetic Retinopathy Screening.
Inventors include S. Priyadharshini; and Dr. S. Kanaga Suba Raja.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: Diabetic Retinopathy (DR) is a progressive retinal disease caused by long-term diabetes and is one of the leading causes of vision impairment and blindness worldwide. Early and accurate detection of DR stages is essential for preventing severe vision loss through timely medical intervention. The present invention proposes an Artificial Intelligence (AI) based automated system for detecting and grading diabetic retinopathy from retinal fundus images using an advanced deep learning framework. In the proposed system, retinal fundus images are first acquired from imaging devices or medical datasets and subjected to pre-processing operations such as noise reduction, contrast enhancement, illumination normalization, and image resizing to improve visual quality and emphasize pathological features. The pre-processed images are then provided to a Compressed Vision Transformer (CVT) based deep learning model, which efficiently extracts discriminative retinal features while reducing computational complexity and model size. The compressed transformer architecture preserves the ability to capture global contextual relationships and fine-grained lesion patterns, while enabling faster inference suitable for large-scale screening applications. The extracted feature representations are subsequently processed by a classification module that categorizes retinal images into multiple diabetic retinopathy severity stages, including normal, mild, moderate, severe, and proliferative DR.
Disclaimer: Curated by HT Syndication.
