Classification of fundus eye images using support vector machines for supporting glaucoma diagnosis

Katarzyna Stąpor, Adrian Brückner


In this paper the new method for automatic classification of fundus eye images into normal and glaucomatous ones is proposed. The cup region is automatically segmented from fundus eye images taken from classical fundus camera. The proposed method makes use of support vector machines classifier with Gaussian kernel. The mean sensitivity is 85 %, while specificity 90%.


SVM classification; MLP classification; glaucoma

Full Text:



Arabas J.: Lectures on genetic algorithms.WNT, Warsaw 2001.

Bezdek J. C.: Pattern recognition with fuzzy objective function algorithms.Plenum Press, New York 1982.

Bishop C. M.: Neural networks for pattern recognition. Clarendon Press, Oxford 1995.

Fletcher R.: Practical methods of optimization. Wiley-Interscience, Chichester 1987.

Gonzalez R.C., Woods R.E.: Digital image processing. Prentice-Hall, 2002.

Guyon I. et al: Structural risk minimization for character recognition. Advances in Neural Information Processing Systems, Nr 4, 1992, 471-479.

Kanski J. et al. Glaucoma: a color manual of diagnosis and treatment. Butterworth-Heinemann, 1996.

Metz M.: Basic principles of ROC analysis. Seminars in Nuclear Medicine, vol. III, nr 4, 1978, 283-298.

Mueller K. R. et al.: An introduction to kernel-based learning algorithms, IEEE Trans. Neural Networks, v.12, Nr 2, 2001, 181-201.

Stąpor K, Brueckner A.: Segmentation of fundus eye images using fuzzy clustering for supporting glaucoma diagnosing. Studia Informatica, submitted for publication.

Stąpor K. et al.: Genetic feature subset selection for classification of eye-cup region in fundus eye images. Studia Informatica, v. 24, Nr 4, 2003, 332-344.

Vapnik V.: The nature of statistical learning theory. Springer Verlag, New York 1995.