Prediction of Cardiovascular Diseases with Retinal Images using Deep Learning

Abstract

Cardiovascular disease (CVD) is the main cause of death globally. The prompt identification and precise diagnosis of CVDs serve as vital for successful treatment and better outcomes for patients. Retinal scanning has evolved as a non-invasive and economical approach for predicting CVD. By harnessing the power of deep learning, specifically Convolutional Neural Networks (CNNs) and MobileNet, this project aims to create an efficient, accurate, and affordable tool for predicting CVDs, ultimately contributing to better patient outcomes and reducing the global burden of these diseases. The suggested model takes use of CNN's ability to autonomously acquire key characteristics from retinal pictures, as well as MobileNet's very light architecture for effective installation. A huge collection of retina pictures, comprising both normal people and patients suffering from heart disease, is used during the training and assessment phases of model. The CNN-type topology is created, using MobileNet as its foundation, adding levels to cater to the unique CVD prediction objective. The system acquires knowledge to correctly classify retina pictures as indicating the presence or absence of CVD after many hours of training and tuning tasks. Functionality is evaluated utilizing common measures like accuracy. The created deep learning framework predicts CVDs using retinal pictures with favourable outcomes, suggesting conceivable advantages for prompt identification, risk evaluation, and economical diagnostics.