Abstract

Background and Purpose: According to the International Diabetes Federation (IDF) the total number of people in India who are suffering from diabetes were around 50.8 million in the year 2010, and it will rise to 87.0 million by 2030. Diabetic Retinopathy is a one of the major complications that exhibits because of Type II diabetes. Diabetic Retinopathy causes blindness in the population of age mostly in between 20 to 64 years. In long term diabetic retinopathy blood vessels disturb the normal flow of fluid out the eye and that comes pressure on the eyeball and this may cause damage to nerves that emerge in glaucoma. Early detection of Diabetic retinopathy and treatment can significantly reduce the risk of vision loss.

Analysis Method: The manual diagnosis of Diabetic retinopathy by ophthalmologists takes time, effort and also includes more costs and can be misdiagnosed if computer aided diagnosis systems are not used. Recently, Deep Learning has become one of the most common methods to achieve high performance results in many areas, especially in medical image analysis and classification. This work addresses the problem of prediction of diabetic retinopathy in advance to avoid further complications in the near future. The proposed classifier was built using MobileNet architecture-a lightweight, mobile friendly architecture, which is trained using retinal fundus images from Aptos 2019 challenge dataset.

Findings: The proposed enhanced model gives an accuracy of 95% and precision, recall, f-1 scores are 0.95, 0.98 and 0.97 respectively. Presented results demonstrate that this model achieves promising results and can be deployed as an application for clinical testing. This work attempts to suggest the diabetic retinopathy complications in advance. The intention of the work is to help the practitioners not to replace the ophthalmologist.

Keywords: Diabetic Retinopathy; MobileNet; CNN

References

1. Zainuddin Z., et al. “A neural network approach in predicting the blood glucose level for diabetic patients”. World Academy of Science, Engineering and Technology50 (2009): 981-988.
2. A Pavate and N Ansari. "Risk Prediction of Disease Complications in Type 2 Diabetes Patients Using Soft Computing Techniques". 2015 Fifth International Conference on Advances in Computing and Communications (ICACC), Kochi (2015): 371-375.
3. https://www.auctoresonline.org/journals/endocrinology-and-disorders/archive/173
4. https://diabetesexpo.com/
5. http://www.eyepacs.com/blog/eyepacs-helps-google-rethink-the-future-of-eye-health
6. R Rajalakshmi., et al. "Automated diabetic retinopathy detection in smartphone-based fundus photography using artificial intelligence”. Eye6 (2018): 1138-1144.
7. J Jin., et al. "Flattened convolutional neural networks for feedforward acceleration”. arXiv preprint arXiv (2014): 1412.5474.
8. M Wang., et al. "Factorized convolutional neural networks”. arXiv preprint arXiv (2016): 1608.04337.
9. Pavate A., et al. "Early prediction of five major complications ascends in diabetes mellitus using fuzzy logic”. In: Nayak, J., Abraham, A., Krishna, B.M., Chandra Sekhar, G.T., Das, A.K. (eds.) Soft Computing in Data Analytics. AISC 758 (2019): 759-768.
10. Andrew G Howard., et al. "MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications”. arXiv:1704.04861v1 [cs.CV] (2017).
11. Y Wu and Z Hu. "Recognition of Diabetic Retinopathy Basedon Transfer Learning”. 2019 IEEE 4th International Conference on Cloud Computing and Big Data Analysis (ICCCBDA), Chengdu, China (2019): 398-401.
12. Y Sun and D Zhang. "Diagnosis and Analysis of Diabetic Retinopathy Based on Electronic Health Records". in IEEE Access 7 (2019): 86115-86120.
13. Y Sun. "The Neural Network of One-Dimensional Convolution-An Example of the Diagnosis of Diabetic Retinopathy". in IEEE Access 7 (2019): 69657-69666.
14. Lam C., et al. "Automated Detection of Diabetic Retinopathy using Deep Learning. AMIA Joint Summits on Translational Science proceedings”. AMIA Joint Summits on Translational Science 2017 (2018): 147-155.
15. Xu K., et al. "Deep Convolutional Neural Network-Based Early Automated Detection of Diabetic Retinopathy Using Fundus Image”. Molecules 12 (2017): 2054.
16. Narayanan BN., et al. "Hybrid machine learning architecture for automated detection and grading of retinal images for diabetic retinopathy”. Journal of Medical Imaging (Bellingham)3 (2020): 034501.
17. Hacisalihoglu RE., et al. "Deep Learning Frameworks for Diabetic Retinopathy Detection with Smartphone-based Retinal Imaging Systems”. Pattern Recognition Letters 135 (2020): 409-417.
18. Riaz H., et al. "Deep and Densely Connected Networks for Classification of Diabetic Retinopathy”. Diagnostics (Basel)1 (2020): 24.
19. Hatanaka Y. “Retinopathy Analysis Based on Deep Convolution Neural Network”. Advances in Experimental Medicine and Biology 1213 (2020): 107-120.
20. https://www.kaggle.com/c/aptos2019-blindness-detection

Citation

Citation: Aruna Pavate., et al. “Diabetic Retinopathy Detection-MobileNet Binary Classifier". Acta Scientific Medical Sciences 4.12 (2020): 86-91.

Copyright

Copyright: © 2020 Aruna Pavate., et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.