Title :
Texture analysis of retinal images to determine nerve fibre loss
Author :
Yogesan, K. ; Eikelboom, R.H. ; Barry, C.J.
Author_Institution :
Centre for Ophalmology & Visual Sci., Western Australia Univ., Nedlands, WA, Australia
Abstract :
The striations of the retinal nerve fiber layer can be recorded photographically. Severe loss of these fibers due to glaucoma can be identified, although this is more difficult with early and diffuse loss. To determine texture analysis is a useful analysis tool to measure nerve fiber loss, five normal and five eyes with glaucomatous nerve fiber loss were analyzed using texture estimator gray level run length matrix. Results show that a number of texture features were able to correctly identify nerve fiber layer loss with an accuracy of 80% to 90%. This study forms the basis for further work on a larger sample of subjects
Keywords :
biomedical imaging; eye; feature extraction; image texture; medical image processing; neurophysiology; photographic applications; glaucoma; nerve fibre loss; retinal images; striations; texture analysis; texture estimator gray level run length matrix; Australia; Biomedical optical imaging; Data mining; Digital images; Electrical capacitance tomography; Image analysis; Image texture analysis; Length measurement; Nerve fibers; Retina;
Conference_Titel :
Pattern Recognition, 1998. Proceedings. Fourteenth International Conference on
Conference_Location :
Brisbane, Qld.
Print_ISBN :
0-8186-8512-3
DOI :
10.1109/ICPR.1998.712039