In this paper, the relation between retinal vessel diameters and glaucoma by means of advanced image processing techniques is examined. Eighteen (18) patients with ocular hypertension who converted to early glaucoma (converters) and nineteen (19) patients with ocular hypertension who remain stable (non-converters) were imaged with a fundus camera at regular intervals. The first acquired image serves as the reference image, which is compared to each follow-up image. The proposed methodology involves the extraction of vessel centerlines by means of differential calculus, the geometrical alignment (registration) of the images using the chamfer matching algorithm, the estimation of vessel diameters by fitting a Gaussian function to intensity profiles. The statistical analysis of the variations of the vessels diameters revealed a significant difference between the two populations. This difference indicates a probable association between glaucoma and variations of the vessel diameter.