AccScience Publishing / AIH / Online First / DOI: 10.36922/aih.2783
ORIGINAL RESEARCH ARTICLE

Deep learning-powered segmentation and classification of diabetic retinopathy for enhanced diagnostic precision

Manoj Saligrama Harisha1 Arya Arun Bhosale1* M. Narender1
Show Less
1 Department of Computer Science and Engineering, The National Institute of Engineering, Mysuru, Karnataka, India
Submitted: 19 January 2024 | Accepted: 1 April 2024 | Published: 6 September 2024
© 2024 by the Author(s). This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution 4.0 International License ( https://creativecommons.org/licenses/by/4.0/ )
Abstract

This study addresses the critical challenge of diabetic retinopathy (DR), a severe complication of diabetes that potentially leads to blindness. We introduce a novel approach to DR detection using transfer learning, leveraging a single fundus photograph to automatically identify the disease’s stage. DR progresses through four stages, posing challenges for early detection, with existing methods often inefficient and prone to disagreements among clinicians. The proposed approach demonstrated in the APTOS 2019 Blindness Detection Competition employs convolutional neural networks (CNNs) and achieved a high quadratic weighted kappa score of 0.92546, highlighting its effectiveness in automatic DR detection and emphasizing the need for timely intervention. This paper first reviews related work, spanning classical computer vision methods to deep learning approaches, with a focus on CNNs. Transfer learning with CNN architectures is explored, showcasing promising results from various studies. Identifying two critical gaps in existing literature, the research emphasizes the need for comprehensive exploration into integrating pre-trained large language models (LLMs) with segmented image inputs for generating test/treatment recommendations. In addition, understanding the dynamic interactions among integrated components, including lesion segmentations, disease classification, and LLMs within web applications, remains essential. The objectives of the study include developing a comprehensive DR detection methodology, exploring and implementing model integration, evaluating performance through competition ranking, contributing significantly to DR detection methodologies, and identifying research gaps. The study encompasses revolutionizing DR detection by integrating cutting-edge technologies, focusing on transfer learning and various model integrations within web applications. The methodology covers data pre-processing, augmentation, segmentation using U-Net neural network architecture, and a detailed training process. The U-Net model demonstrates efficient segmentation of retinal structures with high accuracy and an impressive frames-per-second rate. The results highlight the model’s effectiveness in segmenting blood vessels, hard exudates, soft exudates, hemorrhages, microaneurysms, and the optical disc, with high Jaccard, F1, recall, precision, and accuracy scores. These findings underscore the model’s potential to enhance diagnostic capabilities in retinal pathology assessment, promising improved patient outcomes through timely diagnosis and intervention in combating DR.

Keywords
Diabetic retinopathy
Deep learning
Segmentation
Transfer learning
Convolutional neural networks
Lesion segmentations
Disease classification
U-Net architecture
Funding
None.
Conflict of interest
The authors declare that they have no competing interests.
References
  1. NCHS. Eye Disorders and Vision Loss among U.S. Adults Aged 45 and Over with Diagnosed Diabetes; 2019. Available from: https://stacks.cdc.gov/view/cdc/80081 [Last accessed on 2024 Jul 11].

 

  1. Pezzullo L, Streatfeild J, Simkiss P, Shickle D. The economic impact of sight loss and blindness in the UK adult population. BMC Health Serv Res. 2018;18:63. doi: 10.1186/s12913-018-2836-0.

 

  1. Hann CE, Chase JG, Revie JA, Hewett D, Shaw GM. Diabetic retinopathy screening using computer vision. IFAC Proc Vol. 2009;42:298-303. doi: 10.3182/20090812-3-DK-2006.0086

 

  1. Rohan TE, Frost CD, Wald NJ. Prevention of blindness by screening for diabetic retinopathy: A quantitative assessment. BMJ. 1989;299:1198-1201. doi: 10.1136/bmj.299.6709.1198

 

  1. SNEC. Singapore’s Eye Health; 2019. Available from: https:// www.snec.com.sg [Last accessed on 2024 Jul 11].

 

  1. Priya R, Aruna P. SVM and Neural Network Based Diagnosis of Diabetic Retinopathy; 2012. Available from: https://www. researchgate.net/publication/261177114_svm_and_neural_ network_based_diagnosis_of_diabetic_retinopathy [Last accessed on 2024 Jul 11].

 

  1. Conde P, De la Calleja J, Medina M, Benitez-Ruiz AB. Application of Machine Learning to Classify Diabetic Retinopathy; 2012. Available from: https://www.researchgate. net/publication/259251279_application_of_machine_ learning_to_classify_diabetic_retinopathy [Last accessed on 2024 Jul 11].

 

  1. Pratt H, Coenen F, Broadbent DM, Harding SP, Zheng Y. Convolutional neural networks for diabetic retinopathy. Procedia Comput Sci. 2016;90:200-205. doi: 10.1109/IJCNN.2016.7727515

 

  1. Lam C, Yi D, Guo M, Lindsey T. Automated detection of diabetic retinopathy using deep learning. AMIA Jt Summits Transl Sci Proc. 2018;2017:147-155.

 

  1. Li YH, Yeh NN, Chen SJ, Chung YC. Computer-assisted diagnosis for diabetic retinopathy based on fundus images using deep convolutional neural network. Mobile Inf Syst. 2019;2019:1-14. doi: 10.3390/app9030448

 

  1. Asiri N, Hussain M, Aboalsamh HA. Deep learning based computer-aided diagnosis systems for diabetic retinopathy: A survey. Artif Intell Med. 2019;99:101701. doi: 10.1016/j.artmed.2019.07.009

 

  1. Hagos MT, Kant S. Transfer Learning-based Detection of Diabetic Retinopathy from Small Dataset [Preprint]; 2019. doi: 10.48550/arXiv.1905.07203

 

  1. Sarki R, Michalska S, Ahmed KE, Wang H, Zhang Y. Convolutional Neural Networks for Mild Diabetic Retinopathy Detection: An Experimental Study EyePACs (2015); 2019. Available from: https://www.eyepacs.org [Last accessed on 2024 Jul 11].

 

  1. Fu Y, Zhang G, Lu X, Wu H, Zhang D. U-net: Hard Exudate segmentation for retinal fundus images, expert systems with applications. Expert Syst Appl. 2023;234:120987. doi: 10.1016/j.eswa.2023.120987

 

  1. Fu Y, Zhang G, Li J, Pan D, Wang Y, Zhang D. Fovea localization by blood vessel vector in abnormal fundus images. Pattern Recognit. 2022;129:108711. doi: 10.1016/j.patcog.2022.108711

 

  1. Sahasrabuddhe V, Porwal P, Meriaudeau F, et al. Indian Diabetic Retinopathy Image Dataset (IDRID); 2018. Available from: https://idrid.grand-challenge.org [Last accessed on 2024 Jul 11].

 

  1. Devries T, Taylor GW. Improved Regularization of Convolutional Neural Networks with Cutout [Preprint]; 2017. doi: 10.48550/arXiv.1708.04552

 

  1. Architecture for Computer Vision. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Available from: https://www.cv-foundation.org [Last accessed on 2024 Jul 11].

 

  1. Google Brain. Messidor-2 Diabetic Retinopathy Grades; 2019; Available from: https://research.google/tools/datasets/ messidor2

 

  1. APTOS. APTOS 2019 Blindness Detection; 2019. Available from: https://www.kaggle.com/c/aptos2019-blindness-detection [Last accessed on 2024 Jul 11].

 

  1. Buslaev A, Parinov A, Khvedchenya E, Iglovikov VI, Kalinin AA. Albumentations: Fast and Flexible Image Augmentations [Preprint]; 2020. doi: 10.48550/arXiv.1809.06839

 

  1. Cheng J. A Neural Network Approach to Ordinal Regression [Preprint]; 2007. doi: 10.48550/arXiv.0704.1028

 

  1. Abràmoff MD, Reinhardt JM, Russell SR, et al. Automated early detection of diabetic retinopathy. Ophthalmology. 2010;117:1147-1154. doi: 10.1007/978-3-642-18551-8_17

 

  1. Naveed H, Khan AU, Qiu S, et al. A Comprehensive Overview of Large Language Models [Preprint]; 2023. doi: 10.48550/arXiv.2303.02171

 

  1. Lundberg SM, Lee SI. A Unified Approach to Interpreting Model Predictions. In Advances in Neural Information Processing Systems; 2017. Available from: https://papers.nips.cc/paper_files/paper/2017/ hash/8a20a8621978632d76c43dfd28b67767-abstract.html

 

  1. Silberman N, Ahlrich K, Fergus R, Subramanian L. Case for Automated Detection of Diabetic Retinopathy. In: Artificial Intelligence for Development - Papers from the AAAI Spring Symposium, Technical Report; 2010. doi: 10.1109/CVPR.2010.5540189

 

  1. Paszke A, Gross S, Chintala S, et al. Automatic Differentiation in PyTorch. In NIPS Autodiff Workshop; 2017. Available from: https://openreview.net/forum?id=bjjsrmfcz [Last accessed on 2024 Jul 11].

 

  1. Wan S, Liang Y, Zhang Y. Deep convolutional neural networks for diabetic retinopathy detection by image classification. Comput Electr Eng. 2018;72(10):274-282. doi: 10.1109/ACCESS.2018.2883722

 

  1. Tan M, Le QV. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks [Preprint]; 2019. doi: 10.48550/arXiv.1905.11946

 

  1. Hu J, Shen L, Sun G. Squeeze-and-Excitation Networks [Preprint]; 2017. doi: 10.48550/arXiv.1709.01507

 

  1. Loshchilov I, Hutter F. SGDR: Stochastic Gradient Descent with Restarts [Preprint]; 2016. doi: 10.48550/arXiv.1608.03983

 

  1. Nichol A, Achiam J, Schulman J. On First-order Meta-learning Algorithms [Preprint]; 2018. doi: 10.48550/arXiv.1803.02999

 

  1. Yoon B. A Machine Learning Approach for Efficient Multi-dimensional Integration. Scientific Reports; 2021. Available from: https://www.nature.com/articles/s41598-021-81994-8

 

  1. Krogh A, Hertz JA. A Simple Weight Decay can Improve Generalization. In: Advances in Neural Information Processing Systems; 1992. Available from: https://papers. nips.cc/paper/1992/file/8eefcfdf5990e441f0fb6f3fad709e21- paper.pdf [Last accessed on 2024 Jul 11].

 

  1. Krause J, Gulshan V, Rahimy E, et al. Grader Variability and the Importance of Reference Standards for Evaluating Machine Learning Models for Diabetic Retinopathy [Preprint]; 2017. doi: 10.48550/arXiv.1710.01711

 

  1. Fu Y, Chen J, Li J, Pan D, Yue X, Zhu Y. Optic disc segmentation by U-Net and probability bubble in abnormal fundus images. Pattern Recognit. 2021;117(12):107971. doi: 10.1016/j.patcog.2021.107978

 

  1. Zhu W, Qiu P, Chen X, et al. nnMobile Net: Rethinking CNN for Retinopathy Research [Preprint]; 2024. doi: 10.48550/arXiv.2306.01289

 

  1. Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation. Papers; 2015.
Share
Back to top
Artificial Intelligence in Health, Electronic ISSN: 3029-2387 Print ISSN: 3041-0894, Published by AccScience Publishing