EYES TO THE HEART: PREDICTING CARDIOVASCULAR DISEASE WITH RETINAL IMAGES

Abstract

CVDs are seen as a major burden on global health, and thus there is need for new approaches to early detection and risk assessment. Retinal imaging has become a promising tool for CVD risk prediction due to its ability to assess microvascular alterations linking with systemic vascular pathology without being invasive. This project aims at using retinal images to build a robust predictive model on cardiovascular diseases.

The first stage of the project involves assembling retinal image datasets from diverse populations to ensure that different demographic and clinical factors are represented. These images are then subjected to rigorous preprocessing steps to enhance quality and remove artifacts, thus ensuring the reliability of input data. Advanced image analysis techniques enable extraction of quantitative features related to retinal vascular morphology including vessel diameter, tortuosity, and branching patterns.