AI Based Fundus Image classification Using Bayesian Classifier

Diabetic Retinopathy is a degenerative ocular condition resulting from elevated levels of insulin in the bloodstream, which, if not promptly identified, may lead to vision loss. Figure 1 displays the primary components of digital retinal pictures used for the automated identification of diabetic retinopathy (DR), including blood vessels, optic disc, macula, and fovea [1] . Since the retina is a vascular structure, any alteration in the structure of the retina or blood vessels indicates an abnormality. Therefore, these primary components may be used to identify the existence of any lesion in the retina. Early-stage screening of Fundus Images is crucial to prevent additional damage to the retina. Prompt diagnosis of Proliferative Diabetic Retinopathy (PDR) [6] is of utmost significance and may be achieved by precise identification of blood vessels. The retinal picture may be taken from four different perspectives: (1) Posterior - towards the middle and front, (2) Inferior - towards the higher and front, (3) Superior - towards the lower and front, and (4) Temporal - towards the side. These perspectives may be used to determine the precise position of PDR on the retina. This programme seeks to identify the perspectives that are impacted. This information is beneficial for doctors when recommending laser therapy to their patients.