Hypertension Detection Using AI and Retinal Image Analysis

This research paper introduces a method for detecting hypertensive retinopathy using advanced deep learning techniques, specifically Convolutional Neural Networks (CNNs) applied to retinal fundus images. By analysing a set of over 5469 images from the UK Biobank we were able to achieve an accuracy rate of 81% in identifying signs of hypertensive retinopathy. Additionally, we improved our model by incorporating the detection of multiple eye conditions such, as diabetes, glaucoma, cataracts and age-related macular degeneration leading to overall diagnostic precision and providing a comprehensive screening tool. To make it easier for healthcare professionals to use this model in practice we developed a user-friendly mobile application that allows them to upload retinal images and receive diagnostic predictions displayed in visual formats like histograms and bar charts. This innovation not only makes the model accessible but also provides intuitive visual feedback, aiding in clinical decision-making.