TY  - JOUR
AU  - Tewari, Yogesh
AU  - Parihar, Nitin Singh
AU  - Rautela, Karan
AU  - Kaundal, Nishant
AU  - Diwakar, Manoj
AU  - Pandey, Neeraj Kumar
PY  - 2025
DA  - 2025/06/03
TI  - Diabetic Retinopathy Detection and Analysis with Convolutional Neural Networks and Vision Transformer
JO  - Biomedical Informatics and Smart Healthcare
T2  - Biomedical Informatics and Smart Healthcare
JF  - Biomedical Informatics and Smart Healthcare
VL  - 1
IS  - 1
SP  - 18
EP  - 26
DO  - 10.62762/BISH.2025.724307
UR  - https://www.icck.org/article/abs/BISH.2025.724307
KW  - diabetic retinopathy
KW  - CNN
KW  - ViT
KW  - deep learning
KW  - image classification
AB  - Diabetic Retinopathy occurs when elevated blood sugar levels damage retinal blood vessels, potentially leading to vision impairment. In this paper, we have tested the performance of CNN, ViT and their hybrid models. The dataset used is publicly available on Kaggle and the dataset contained around 35,000 retinal images which were divided into 5 classes namely No DR, Mild DR, Moderate DR, Severe DR and Proliferative DR. In CNN we tested 4 different architectures in which we achieved the best accuracy of 75.4% with Resnet50 architecture and with ViT model we achieved an accuracy of 83.9% and from the hybrid model we achieved an accuracy of 88.4% from the Resnet50 + ViT. The results shown by the models were promising but there were some gaps in the study. The dataset used was skewed towards NO DR class. For future work more balanced datasets with some data augmentation techniques could be used. Additionally, the study used only 50 epochs which can be increased in future work to use the model to their full potential.
SN  - 3068-5524
PB  - Institute of Central Computation and Knowledge
LA  - English
ER  - 

@article{Tewari2025Diabetic,
  author = {Yogesh Tewari and Nitin Singh Parihar and Karan Rautela and Nishant Kaundal and Manoj Diwakar and Neeraj Kumar Pandey},
  title = {Diabetic Retinopathy Detection and Analysis with Convolutional Neural Networks and Vision Transformer},
  journal = {Biomedical Informatics and Smart Healthcare},
  year = {2025},
  volume = {1},
  number = {1},
  pages = {18-26},
  doi = {10.62762/BISH.2025.724307},
  url = {https://www.icck.org/article/abs/BISH.2025.724307},
  abstract = {Diabetic Retinopathy occurs when elevated blood sugar levels damage retinal blood vessels, potentially leading to vision impairment. In this paper, we have tested the performance of CNN, ViT and their hybrid models. The dataset used is publicly available on Kaggle and the dataset contained around 35,000 retinal images which were divided into 5 classes namely No DR, Mild DR, Moderate DR, Severe DR and Proliferative DR. In CNN we tested 4 different architectures in which we achieved the best accuracy of 75.4\% with Resnet50 architecture and with ViT model we achieved an accuracy of 83.9\% and from the hybrid model we achieved an accuracy of 88.4\% from the Resnet50 + ViT. The results shown by the models were promising but there were some gaps in the study. The dataset used was skewed towards NO DR class. For future work more balanced datasets with some data augmentation techniques could be used. Additionally, the study used only 50 epochs which can be increased in future work to use the model to their full potential.},
  keywords = {diabetic retinopathy, CNN, ViT, deep learning, image classification},
  issn = {3068-5524},
  publisher = {Institute of Central Computation and Knowledge}
}