<p>Diabetic Retinopathy (DR) is a disease of the retina caused by diabetes mellitus that threatens vision. DR can progress without giving any signs until it causes sudden loss of vision. So, early detection of DR is crucial. Manual diagnosis of DR demands a high level of expertise and efforts from trained ophthalmologists, which is costly, time-consuming, and prone to misdiagnosis. The availability of trained ophthalmologists is significantly less, so an automated approach is required to efficiently detect the disease in its early stages with better diagnosis and less expense. Computer-Aided-Diagnosis (CAD) systems play a crucial role in assisting ophthalmologists and retinal experts in early disease detection. Using computer vision, image processing, and artificial intelligence techniques, including machine learning and deep learning, has shown significant promise in medical imaging for developing automated diagnostic tools. This paper reviews various datasets of retinal fundus images and explores corresponding methodologies for DR detection. Over the past few years, numerous studies have focused on DR detection, encompassing retinal image segmentation, classification, and detection. This paper provides a comprehensive analysis of these studies, identifying gaps and challenges while offering insights into future directions for improving DR detection using retinal fundus imaging. Further, it emphasizes real-world deployment challenges such as interpretability, generalizability, and integration into clinical workflows.</p>