Laying eyes on glaucoma
An obstruction of the natural flow of aqueous humor may result in an increased pressure in the eyeball that may eventually damage the optic disc and harm vision. The so-called glaucoma disease, which is one of the most frequent reasons for blindness may be effectively treated if it is early detected, Taking advantage of an existing optical databank, the GLAUCAD project developed various software tools for supporting clinicians in fast and reliable glaucoma diagnosis. As such, a new software tool exploits retinal images that picture the eye part that transmits light and images to the brain through the optic nerve. The basic concept behind this tool is to identify vessel deformations if any in order to specify abnormalities in the subject's vision. Thereby, it supports automatic glaucoma detection, in presence of vessel movement due to pathological or non-pathological causes. This software results in the classification of subjects into two types, which are converter/non-converter that present abnormal/normal vessel deformations over time. Two pair of retinal images acquired with a few months interval are first registered and then subjected to analysis. This involves identification of the central lines of retinal vessels and calculation of shape descriptors for user-selected vessels. If the shape descriptors for user-selected vessels display different values for the two time-sequenced images then the subject is classified as a converter type. The software has been proven very accurate both in terms of derived results and image registration for animation generation. The latter constitutes a very popular method for routine medical inspections. Retinal image pairs are processed in batches and all key vessels in a retinal image pairs are subjected to examination. The tool may be either used as a stand-alone device or offered through a service network.
