Ophthalmic diseases such as glaucoma are associated with progressive changes in the structure of the optic disc (OD) and parapapillary atrophy (PPA). These structural changes may therefore have relevance to other systemic diseases. The size and location of OD and PPA can be used as registration landmarks for monitoring changes in features of the fundus of the eye. Retinal vessel evaluation, for example, can be used as a biomarker for the effects of multiple systemic diseases, or co-morbidities. This thesis presents the first computer-aided measuring tool that detects and quantifies the progression of PPA automatically on a 2D retinal fundus image in the presence of image noise. An automated segmentation system is described that can detect features of the optic nerve. Three novel approaches are explored that extract the PPA and OD region approximately from a 2D fundus image. The OD region is segmented using (i) a combination of active contour and morphological operations, (ii) a modified Chan-Vese algorithm and (iii) a combination of edge detection and ellipse fitting methods. The PPA region is identified from the presence of bright pixels in the temporal zone of the OD, and segmented using a sequence of techniques, including a modified Chan-Vese approach, thresholding, scanning filter and multi-seed region growing methods. The work demonstrates for the first time how the OD and PPA regions can be identified and quantified from 2D fundus images using a standard fundus camera.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.