Diabetic Retinopathy Using Deep Learning

Dr. Virupakshappa; Almas1

1

Publication Date: 2024/08/22

Abstract: Untreated diabetic retinopathy, a complication of uncontrolled diabetes, may lead to total blindness if not addressed promptly. Consequently, in order to avoid the serious complications of diabetic retinopathy, it is crucial to diagnose the condition early and treat it medically. Patients go through a lot of pain and suffering as ophthalmologists manually identify diabetic retinopathy. With the use of an automated method, diabetic retinopathy may be detected more rapidly, allowing for easier follow-up therapy to prevent more eye damage. This paper presents a machine learning strategy for feature extraction including exudates, hemorrhages, and micro aneurysms. The strategy involves a hybrid classifier that integrates support vector machine, k closest neighbour, random forest, logistic regression, and multilayer perceptron networks. To further assist in DR stage image recognition, for instance to detect blood vessels, future research may center on applying object identification techniques based on convolutional neural networks (CNNs).

Keywords: Diabetic, Deep learning, Retinopathy.

DOI: https://doi.org/10.38124/ijisrt/IJISRT24AUG158

PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT24AUG158.pdf

REFERENCES

  1. Thien M. Ha, Matthias Zimmermann, Horst Bunke, Off-line handwritten numeral string recognition by combining segmentation based and segmentation-free methods, Pattern Recognition, Volume 31, Issue 3, 1998, Pages 257-272, ISSN 0031-3203
  2. Zhan H., Wang Q., Lu Y. (2017) Handwritten Digit String Recognition by Combination of Residual Network and RNN-CTC. In: Liu D., Xie S., Li Y., Zhao D., El-Alfy ES. (eds) Neural Information Processing. ICONIP 2017. Lecture Notes in Computer Sci-ence, vol 10639. Springer, Cham. https://doi.org/10.1007/978-3-319-70136- 3_62
  3. Filatov, A. Gitis and I. Kil, "Graph-based handwritten digit string recognition," Pro-ceedings of 3rd International Conference on Document Analysis and Recognition, 1995, pp. 845-848 vol.2, doi: 10.1109/ICDAR.1995.602033.
  4. H. Zhan, S. Lyu and Y. Lu, "Handwritten Digit String Recognition using Convolutional Neural Network," 2018 24th InternationalConference on Pattern Recognition (ICPR), 2018, pp. 3729-3734, doi: 10.1109/ICPR.2018.8546100.
  5. L. S. Oliveira and R. Sabourin, "Support vector machines for handwritten numerical string recognition," Ninth International Workshop on Frontiers in Handwriting Recogni-tion, 2004, pp. 39-44, doi: 10.1109/IWFHR.2004.99.
  6. C. Zanchettin, B. L. D. Bezerra and W. W. Azevedo, "A KNN-SVM hybrid model for cursive handwriting recognition," The 2012 International Joint Conference on Neural Networks (IJCNN), 2012, pp. 1-8, doi: 10.1109/IJCNN.2012.6252719.
  7. X. -. Zhou, J. -. Yu, C. -. Liu, T. Nagasaki and K. Marukawa, "Online Handwritten Jap-anese Character String Recognition Incorporating Geometric Context," Ninth Interna-tional Conference on Document Analysis and Recognition (ICDAR 2007), 2007, pp. 48-52, doi: 10.1109/ICDAR.2007.4378673.
  8. Ahlawat S, Choudhary A, Nayyar A, Singh S, Yoon B. Improved Handwritten Digit Recognition Using Convolutional Neural Networks (CNN). Sensors.2020;20(12):3344. https://doi.org/10.3390/s20123344.