Diabetic Retinopathy is a severe eye condition originating as a result of long term diabetes mellitus. Timely detection is essential to prevent it from progressing to more advanced stages. Manual detection of DR is labor-intensive and time-consuming, requiring expertise and extensive image analysis. Our research aims to develop a robust and automated deep learning model to assist healthcare professionals by streamlining the detection process and improving diagnostic accuracy. This research proposes a multi-classification framework using Transfer Learning for diabetic retinopathy grading among diabetic patients. An image based dataset, APTOS 2019 Blindness Detection, is utilized for our model training and testing. Our methodology involves three key preprocessing steps: 1) Cropping to remove extraneous background regions, 2) Contrast enhancement using CLAHE (Contrast Limited Adaptive Histogram Equalization) and 3) Resizing to a consistent dimension of 224x224x3. To address class imbalance, we applied SMOTE (Synthetic Minority Over-sampling Technique) for balancing the dataset. Data augmentation techniques such as rotation, zooming, shifting, and brightness adjustment are used to further enhance the model's generalization. The dataset is split to a 70:10:20 ratios for training, validation and testing. For classification, EfficientNetB3 and Xception, two transfer learning models, are used after fine-tuning which includes addition of dense, dropout and fully connected layers. Hyper parameters such as batch size, no. of epochs, optimizer etc were adjusted prioir model training. The performance of our model is evaluated using various performance metrics including accuracy, specificity, sensitivity and others. Results reveal the highest test accuracy of 95.16% on the APTOS dataset for grading diabetic retinopathy into five classes using the EfficientNetB3 model followed by a test accuracy of 92.66% using Xception model. Our top-performing model, EfficientNetB3, was compared against various state-of-the-art approaches, including DenseNet-169, hybrid models, and ResNet-50, where our model outperformed all these methodologies.

Diabetic retinopathy stands as a significant concern for individuals managing diabetes. It is a severe eye condition that targets the delicate blood vessels within the retina. As it advances, it can inflict severe vision impairment or complete blindness in extreme cases. Regular eye examinations are vital for individuals with diabetes to detect abnormalities early. Detection of diabetic retinopathy is challenging and a time-consuming process, but deep learning and transfer learning techniques offer vital support by automating the process, providing accurate predictions, and simplifying diagnostic procedures for healthcare professionals. This study introduces a multi-classification framework for grading diabetic retinopathy into five classes using Transfer Learning and data fusion. The objective is to develop a robust, automated model for diabetic retinopathy detection to enhance the diagnostic process for healthcare professionals. We fused two distinct datasets, APTOS and IDRiD, which resulted in a total of 4178 fundus images. The merged dataset underwent preprocessing to enhance image quality and to remove unwanted regions, noise and artifacts from the fundus images. The pre-processed dataset is then resized and a balancing technique called SMOTE is applied to it due to uneven class distribution present among classes. To increase diversity and size of the dataset, data augmentation techniques including flipping, brightness adjustment and contrast adjustment are applied. The dataset is split into 80:10:10 ratios for training, validation, and testing. Two pre-trained models, EfficientNetB5 and DenseNet121, are fine-tuned and training parameters like batch size, number of epochs, learning rate etc. are adjusted. The results demonstrate the highest test accuracy of 96.06% is achieved by using EfficientNetB5 model followed by 91.40% test accuracy using DenseNet121 model. The performance of our best model i.e. EfficientNetB5, is compared with several state-of-the-art approaches, including DenseNet-169, Hybrid models and ResNet-50 where our model outperformed these methodologies in terms of test accuracy.