ICA analysis of retina images for glaucoma classification

Glaucoma represent one of the most frequent causes of partial loss of the visual field. It comes along with an ongoing destruction of the optic nerve caused by an increased pressure of the eye liquid. The disease becomes obvious from investigations of the retina with scanning laser microscopes. In this report an image analysis and classification system based on independent component analysis and k-nearest-neighbor classification is proposed. The method is tested with 120 selected retina images collected with the Heidelberg Retina Tomograph and achieves a classification rate of 91%.