The Role of Features in Image Retrieval System: A Survey

In recent years, digital images have gained a significant role through the growth of computer technology and multimedia information systems. Image retrieval systems assist in retrieving images similar to query images from vast databases. However, retrieving similar images from a large database remains a challenging task, highlighting the demand for successful image retrieval techniques. The effectiveness of such systems depends on the features and techniques employed for retrieval. In this paper, the existing works on image retrieval systems are analyzed and present a survey addressing three key aspects: features, datasets, and performance metrics. The features used in image retrieval systems are discussed under various categories, including Content Based Image Retrieval (CBIR), Text Based Image Retrieval (TBIR), and Semantic Based Image Retrieval (SBIR) systems. Additionally, the datasets utilized for image retrieval and the performance metrics used to evaluate these systems are also presented.