The eyes have it

A regular eye examination is a solid part of maintaining eye health allowing problems to be identified sooner rather than later. Part of such an examination will commonly involve checking the lens of the eye as well as the interior of the eye. Recording a digital image of the light-sensitive retina at the back of the eye is frequently used by ophthalmologists, optometrist and others to detect vascular disorders, such as diabetic retinopathy, evidence of glaucoma, age-related macular degeneration, and optic neuritis.

One of the issues with using computers to analyse a retinal image is to ensure the precise localisation of the optic disc within the retinal image in the computer so that features can be compared more accurately and problems be detectable by software with fewer false positives or false negatives for pathologies and other concerns. A new Open Access paper, published in the International Journal of Computational Vision and Robotics, describes a JAYA algorithm that performs with 99 per cent accuracy in localising the optic disc within the retinal image It uses a novel fitness function to do so. This new system performs better than other methods previously reported in the scientific literature based on tests with a publicly available database of retinal images.

Around 150 million people are afflicted with glaucoma and diabetic retinopathy. Both conditions can lead to compromised sight and ultimately blindness. Unfortunately, there are no obvious external symptoms of the creeping pathology of either condition. Regular detailed examination of the eye is therefore essential, especially of those people in at-risk groups, to detect the earliest signs of the conditions as soon as possible so that treatment can be started to retard the problem.

Visual examination of retinal images is time-consuming for optometrists and takes a great deal of skill and patience. As with many other kinds of medical examination, there is always room for error or the need for the proverbial second opinion. As such, a computerised algorithmic approach could be a useful tool for optometrists to quickly highlight problems with their patients’ eyes without the labour-intensive screening. Once the computer has highlighted a problem the optometrist can then carry out further visual examination of their own to ascertain the seriousness of any condition that is flagged and then prescribe the requisite course of action for their patient.

Kumar, B.V., Zhang, S., Wu, T., Prakash, J., Zhou, L. and Li, K. (2022) ‘A novel JAYA algorithm for optic disc localisation in eye fundus images‘, Int. J. Computational Vision and Robotics, Vol. 12, No. 3, pp.324-342.