Heart disease prediction using retinal eye images involves analyzing the retina's blood vessels and structures to identify signs of cardiovascular conditions. Advanced imaging techniques, combined with machine learning algorithms. It is done by combining them with health indicators like hypertension Retinopathy, Normal Fundus and diabetes to non-invasively assess cardiovascular risk.
The used dataset is from Kaggle notebook to train the model as training as testing by 90 and 10 percentages respectively. Later the dataset had been undergone preprocessing. The Inception V3 of the CNN classifier is used to classify and extract the required features.
Plotting various graphs for training accuracy, training loss and ROC representation
Confusion matrix determines the classes predicted as true and false positives
By considering all the required classes of retinal structures, the project predicts the status of heart disease.
