A Review on the Analysis of Glaucoma and its Automated Detection Methods from Fundus Images using Machine Learning Techniques

Glaucoma a long-term degenerative disease affecting the optic nerve, which leads to permanent vision impairment if not identified and treated early. As a major contributor to global blindness, it often progresses without symptoms until advanced stages. Conventional diagnostic techniques such as assessing Intraocular pressure (IOP), evaluating the optic disc, and performing visual field tests demand expert interpretation and are subject to variability. Consequently, the detection of Glaucoma using machine learning (ML) methods has become a significant research focus in the field of ophthalmology. This paper presents a survey of current ML-based approaches for detecting glaucoma using Fundus imagery. It covers Preprocessing steps, Feature extraction strategies, and Classification algorithms, offering a comparative analysis of their performance in clinical applications. Notable datasets such as RIM-ONE, ORIGA, DRISHTI-GS1, and REFUGE are discussed inoder to support reproducible research. The paper also addresses key challenges, including class imbalance, diverse image quality, and the transparency of AI models. Furthermore, emerging trends like Transfer learning, Federated learning, and Multi-modal systems are examined for their potential to enhance early detection and assist clinical decision-making.