Prediction of Cardiovascular Diseases with Retinal Images Using Deep Learning

Abstract

Cardiovascular diseases (CVDs) are the leading cause of death worldwide, highlighting the need for timely detection and accurate diagnosis. This research develops a deep learning framework using convolutional neural networks (CNNs) combined with the MobileNet architecture to identify CVDs from retinal images. The model utilizes CNNs to automatically extract important features while benefiting from MobileNet's efficiency. A diverse dataset of retinal images, including both healthy individuals and those with CVDs, is used for training and testing, with preprocessing steps like resizing, normalization, and data augmentation to enhance dataset quality. The tailored CNN architecture effectively distinguishes between retinal images indicating CVD presence or absence, achieving promising accuracy in predictions. This model offers significant advantages for early detection, risk assessment, and cost-effective diagnosis of cardiovascular conditions, potentially aiding healthcare providers in making informed decisions and implementing timely interventions. Further validation and integration into clinical practice are essential to assess its full impact on patient care.