A Study of Diabetic Retinopathy Classification Using Support Vector Machine

  • Authors

    • Nur Izzati Ab Kader
    • Umi Kalsom Yusof
    • Syibrah Naim
    2018-12-09
    https://doi.org/10.14419/ijet.v7i4.31.27823
  • Diabetic Retinopathy, Machine Learning, Classification, Support Vector Machine, Kernel.

  • Diabetic Retinopathy (DR) is a diabetic complication which can cause blindness. As DR cases keep increasing, ophthalmologists are forced to diagnose a large number of retinal images daily. Generally, the diabetic eye screening is done manually using qualitative scale to detect abnormalities on the retina. Although this approach is useful, the detection is not accurate; and create a need for a tool that can help the experts to classify the severity of DR to establish adequate therapy. Previous researchers have studied machine learning to propose an automatic DR classification. However, it needs to be improvised especially in terms of accuracy. Hence, this paper aimed to find classifier with optimal performance in the study of DR classification. This study considered three classes of diabetic patients which were patients who do not have DR (NODR), patients with non-proliferative DR (NPDR) and patients with proliferative DR (PDR), instead of focusing only on two classes (NO DR, DR). Support Vector Machine was used in this research due to the success of many classification problems that had been proposed which produced good result. The results obtained showed that SVM gave the best accuracy, 76.62% with average sensitivity of 0.8081 and average specificity of 0.8376 respectively.

     

     

     

  • References

    1. [1] “Global reports on diabetesâ€, World Health Organization, Tech. Rep.,2016.

      [2] W. M. D. W. Zaki, M. A. Zulkifley, A. Hussain, W. H. W. Halim,N. B. A. Mustafa, and L. S. Ting, “Diabetic retinopathy assessment:Towards an automated systemâ€, Biomedical Signal Processing and Control, vol. 24,2016, pp. 72–82.

      [3] B. Wu, W. Zhu, F. Shi, S. Zhu, and X. Chen, “Automatic detection of microaneurysms in retinal fundus imagesâ€, Computerized Medical Imaging and Graphics, vol. 55, 2017, pp. 106–112.

      [4] P. J. Navarro, D. Alonso, and K. Stathis, “Automatic detection of microaneurysms in diabetic retinopathy fundus images using the l* a*b color spaceâ€, JOSA A, vol. 33, no. 1, 2016, pp. 74–83.

      [5] A. Bourouis, M. Feham, M. A. Hossain, and L. Zhang, “An intelligent mobile based decision support system for retinal disease diagnosisâ€, Decision Support Systems, vol. 59, 2014, pp. 341–350.

      [6] S. Piri, D. Delen, T. Liu, and H. M. Zolbanin, “A data analytics approach to building a clinical decision support system for diabetic retinopathy: Developing and deploying a model ensembleâ€, Decision Support Systems, 2017.

      [7] G. Parthiban and S. Srivatsa, “Comparing naive bayes and decisiontree techniques for predicting the risk of diabetic retinopathyâ€, Digital Signal Processing, vol. 7, no. 5, 2015, pp. 141–145

      [8] H. Evirgen and M.C erkezi, “Prediction and diagnosis of diabetic retinopathy using data mining techniqueâ€, Turkish Online Journal of Science and Technology, vol. 4, no. 3, 2014.

      [9] V. Balakrishnan, M. R. Shakouri, H. Hoodehet al., “Predictions using data mining and case-based reasoning: a case study for retinopathy,†World Academy of Science, Engineering and Technology, International Journal of Medical, Health, Biomedical, Bioengineering and Pharma- ceutical Engineering, vol. 6, no. 3, 2012, pp. 55–58.

      [10] M. Skevofilakas, K. Zarkogianni, B. G. Karamanos, and K. S. Nikita,“A hybrid decision support system for the risk assessment of retinopathy development as a long term complication of type 1 diabetes mellitusâ€, IEEE, 2010, pp. 6713–6716.

      [11] H.-Y. Huang and C.-J. Lin, “Linear and kernel classification: When touse which?â€, inProceedings of the 2016 SIAM International Confer- ence on Data Mining, 2016, pp. 216–224.

      [12] J. Weston and C. Watkins, “Multi-class support vector machinesâ€,Cite-seer, 1998.

      [13] V. Franc and V. Hlav ac, “Multi-class support vector machine,â€, IEEE, vol. 2, 2002, pp. 236–239.

      [14] Z. Wang and X. Xue, “Multi-class support vector machineâ€, Springer, 2014, pp. 23–48.

      [15] S. Knerr, L. Personnaz, and G. Dreyfus, “Single-layer learning revis- ited: a stepwise procedure for building and training a neural networkâ€, in Neurocomputing. Springer, 1990, pp. 41–50

      [16] J. Friedman, “Another approach to polychotomous classificationâ€, Technical report, Department of Statistics, Stanford University, Tech. Rep.,1996.

      [17] K. Polat, S. G unes , and A. Arslan, “A cascade learning system for clas- sification of diabetes disease: Generalized discriminant analysis and least square support vector machineâ€, Expert systems with applications, vol. 34, no. 1, 2008, pp. 482–487.

      [18] C.-H. Weng, T. C.-K. Huang, and R.-P. Han, “Disease predictionwith different types of neural network classifiersâ€, Telematics andInformat- ics, vol. 33, no. 2, 2016, pp. 277–292

      [19] L. M.Manevitz and M. Yousef, “One-class svms for documet classifi- cationâ€, Journal of Machine Learning Research, pp. 139–154, 2001.

      [20] M.L. McHugh, “Interrater reliability: the kappa statisticâ€, Biochemi- amedical, vol. 22, no. 3, 2012, pp. 276–282.

  • Downloads

  • How to Cite

    Izzati Ab Kader, N., Kalsom Yusof, U., & Naim, S. (2018). A Study of Diabetic Retinopathy Classification Using Support Vector Machine. International Journal of Engineering & Technology, 7(4.31), 521-527. https://doi.org/10.14419/ijet.v7i4.31.27823