Severity Differentiation and Detection of Glaucoma using Pulikulam Cattle Optimization Algorithm (PCOA)-Based CNN in Retinal Fundus Image

Abstract

Glaucoma, a debilitating eye condition, poses a significant threat to vision by damaging optic nerve fibers and astrocytes irreversibly. Early detection of glaucoma is crucial for timely intervention and preservation of vision. Retinal image-based detection methods offer a non-invasive approach for early diagnosis, which can alleviate the burden on ophthalmologists and improve patient outcomes. In this study, we propose a novel method called the Pulikulam Cattle Optimization Algorithm (PCOA) for glaucoma detection using retinal fundus images. The PCOA algorithm is employed to optimize the weights of a Convolutional Neural Network (CNN) classifier, enhancing its accuracy and efficiency in detecting glaucomatous features. The fitness function of the PCOA algorithm aims to minimize error values, leading to the identification of optimal solutions for glaucoma detection. The resulting optic disc area extracted from retinal images serves as a crucial indicator for distinguishing between healthy and glaucomatous eyes. We conducted comprehensive evaluations using diverse datasets, demonstrating well-organized clustering, precise classification, and superior performance compared to existing methods. Our proposed approach achieved accuracy levels exceeding 95%, underscoring its effectiveness in glaucoma detection. The findings of this study contribute to advancing glaucoma detection technology and hold promise for improving clinical outcomes and patient care.