Early Detection and Symptoms of Diabetic Retinopathy in Children: A Deep Learning Approach

Authors

  • Aishwarya G
  • Smitha BA
  • Nivedita G Y
  • Sunitha K
  • Bhagyashree Ambore

Keywords:

Diabetes Retinopathy, Convolutional Neural Networks, Paediatric Ophthalmology, Deep Learning, Early Detection, Fundus Imaging, Sensitivity, AUC, and Specificity

Abstract

Diabetic retinopathy (DR) in children remains the largest single cause of blindness globally with Type 1 diabetes. Early discovery is vital to stop irreversible visual loss. This article explores the application of deep learning, mainly convolutional neural networks (CNNs), to early detection and DR classification in children. A dataset of 10,000 retinal images of paediatric patients was used to train and test a pre-trained ImageNet CNN model that was fine-tuned for the identification of paediatric disease. The model categorizes DR into five groups: no DR, mild NPDR, moderate NPDR, severe NPDR, and proliferative DR. The model has a sensitivity of 88.6% for mild NPDR, a specificity of 89.2% for healthy retinas, and an overall accuracy of 92.3%. Furthermore, the model has an AUC of 0.97, which suggests outstanding discriminative ability. Deep learning methods can be valuable for early detection of DR among young individuals and result in effective therapeutic intervention on time and enhanced visual outcomes.

Downloads

Download data is not yet available.

References

Priyanga P, Satish Kumar S, Bhagyashree Ambore, Sunitha K, Nivedita G Y, Aishwarya G, (2025) Early Detection of Diabetic Retinopathy Using Transfer Learning with VGG16: A Deep Learning Approach for Retinal Fundus Analysis. Journal of Neonatal Surgery, 14 (22s), 651-660.

S. Ramasamy et al., "Transfer learning with deep convolutional neural network for early detection of diabetic retinopathy in children," Comput. Methods Programs Biomed., vol. 190, p. 105362, 2020.

R. Rajalakshmi, R. Subashini, R. Anjana, and V. Mohan, "Automated diabetic retinopathy detection in smartphone-based fundus photography using deep learning for teleophthalmology," JAMA Ophthalmol., vol. 136, no. 10, pp. 1186–1194, 2018.

A. Pratt, F. Coenen, and D. M. Broadbent, "Convolutional neural networks for diabetic retinopathy," Procedia Comput. Sci., vol. 90, pp. 200–205, 2016.

M. Mookiah et al., "Computer-aided diagnosis of diabetic retinopathy: A review," Comput. Biol. Med., vol. 43, no. 12, pp. 2136–2155, 2013.

S. Ramasamy et al., "Transfer learning with deep convolutional neural network for early detection of diabetic retinopathy in children," Comput. Methods Programs Biomed., vol. 190, p. 105362, 2020.

T. Kaur and S. Gandhi, "Automated grading of diabetic retinopathy using deep learning," Health Technol., vol. 11, pp. 693–700, 2021.

N. R. Y. Bassam, "Hybrid CNN-based ensemble model for detecting diabetic retinopathy," Multimed. Tools Appl., vol. 80, no. 9, pp. 13769–13788, 2021.

Y. Yang, H. Yan, and X. Jiang, "A comprehensive ensemble model for diabetic retinopathy detection," IEEE Access, vol. 8, pp. 76547–76556, 2020.

A. Yaqoob, S. S. Bukhari, and M. Shahzad, "Data augmentation and transfer learning for diabetic retinopathy classification using deep CNNs," Biomed. Signal Process. Control, vol. 70, p. 103060, 2021.

Z. Wang, Q. Liu, and X. Dou, "Diabetic retinopathy diagnosis via deep learning and attention mechanisms," J. Med. Syst., vol. 44, no. 9, pp. 1–10, 2020.

A. Chan et al., "Deep learning in pediatric diabetic retinopathy screening: Challenges and solutions," Pediatr. Ophthalmol., vol. 27, pp. 215–223, 2019.

M. Islam, S. Zhang, and P. Ren, "Assessing the generalizability of DR detection models across pediatric datasets," in Proc. IEEE EMBC, 2021, pp. 6351–6355.

J. Xu et al., "Improved pediatric DR classification using contrast enhancement and CNN fine-tuning," Comput. Med. Imaging Graph., vol. 89, p. 101870, 2021.

da Rocha, D.A., Ferreira, F.M.F. & Peixoto, Z.M.A. Diabetic retinopathy classification using VGG16 neural network. Res. Biomed. Eng. 38, 761–772 (2022). https://doi.org/10.1007/s42600-022-00200-8

Downloads

Published

2025-05-20

How to Cite

1.
G A, BA S, G Y N, K S, Ambore B. Early Detection and Symptoms of Diabetic Retinopathy in Children: A Deep Learning Approach. J Neonatal Surg [Internet]. 2025May20 [cited 2025Sep.21];14(24S):817-824. Available from: https://www.jneonatalsurg.com/index.php/jns/article/view/6146

Issue

Vol. 14 No. 24S (2025): Journal of Neonatal Surgery

Section

Original Article

License

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

You are free to:

  • Adapt — remix, transform, and build upon the material for any purpose, even commercially.

Terms:

  • Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
  • No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.