Rider Manta Ray Foraging Optimization-based Generative Adversarial Network and CNN feature for detecting glaucoma

Abstract

Glaucoma is a leading and rapidly increasing disease in recent years, which affects the eye's optic nerve and causes everlasting vision loss. The early recognition of glaucoma is a significant one for decreasing the risk of permanent loss of sight. Therefore, it is necessary to detect glaucoma disease in the initial stage. Moreover, optic cup segmentation from retinal fundus images is an essential procedure for automatic glaucoma detection. In this paper, an efficient glaucoma detection approach is developed using Rider Manta-Ray Foraging Optimization-based General Adversarial Network ((Rider MRFO-based GAN). Here, the optic disc segmentation process is performed by the Fuzzy Local Information C-Means clustering (FLICM clustering). In addition, the sparking process is also employed in this glaucoma detection method for the blood vessel detection process. Several significant features in this detection model, namely mean, standard deviation, variance, kurtosis, skewness, entropy, and CNN features, are extracted to detect glaucomatous images further. The developed Rider MRFO approach is newly developed by the MRFO technique and Rider Optimization Algorithm (ROA). Additionally, the developed glaucoma detection technique performs better based on various parameters, like specificity, sensitivity, and accuracy. Hence, the developed Rider MRFO-based GAN model showed improved results with the highest accuracy of 0.96, the sensitivity of 0.94, and the specificity of 0.89.