Automated Detection of Diabetic Retinopathy Severity usingSelf-supervised Learning and Embedded Knowledge Detection
Dates: 2024
Principal Investigator: Dr. Fahkri Alam Khan
Description: Automating diabetic retinopathy (DR) screening is crucial for early detection and resource saving, especially in Saudi Arabia with rising diabetes rates and a shortage of ophthalmologists. Traditional methods face challenges due to the time-consuming assessment of small disease biomarkers. This project aims to address these obstacles through four key tasks:
1. Research on automatic recognition of DR severity in fundus images to streamline screening.
2. Evaluation of single-view versus dual-view fundus images for DR detection, utilizing appropriate datasets.
3. Exploration of computer vision and deep learning techniques, including self-supervised learning, for automatic DR recognition.
4. Validation of the proposed framework through collaboration with ophthalmologists.
By leveraging AI technologies, this project seeks to expedite the diagnostic process, cater to a larger patient population, and ultimately prevent vision loss among diabetic individuals.