A Review of Image Retrieval Methods: Progress from Feature-Based to Deep Learning Approaches

Image retrieval systems are essential for efficiently accessing relevant visual content from massive datasets. Over the years, retrieval methods have advanced significantly, transitioning from simple keyword-based systems to content-based models and, more recently, to deep learning-based approaches. This review outlines major categories of image retrieval techniques, including text-based retrieval, content-based image retrieval (CBIR), machine learning-enhanced methods, and current trends in deep learning and hybrid frameworks. The paper also discusses their respective strengths, limitations, and prospects for further research.