Neural network classification of eye-cup region in fundus eye images

Katarzyna Stąpor, Marek Rzendkowski, Michał Mazurkiewicz


In this paper the new method for classification of digital fundus eye images into normal and glaucomatous ones is proposed. The classifier used is based on multi-layer perceptron. The performance of the classifier is 94,77%. The obtained results arę encouraging.


neural network classification; backpropagation rule; glaucoma