Diabetic-Retinopathy-Disease-Grading-Deep-Learning-Model

The purpose of this deep learning model is to predict the risk of diabetic retinopathy from low severity 0 to high severity 4 from the retinal image.

1. Implemented in python programming language.

2. Extracted features using pretrained model VGG16.

3. Trained using LightGBM algorithm.

4. Used class weighting technique to handle imbalanced diabetic risk levels.

5. Evaluated the performance with the help of confusion matrix,accuracy,precision,recall and f1 score.