Image Indexing Approaches for Enhanced Content-Based Image Retrieval: An Overview

Abstract

The effective management and analysis of vast collections of digital images rely heavily on image indexing. This process facilitates image access through the creation of a textual description or digital signature using metadata and visual data such as color, shape, and texture. Manual and automatic indexing are two commonly used techniques for generating image indexing. This article explores the three main approaches used by current image retrieval systems, namely text-based, content-based, and ontology-based approaches. Specifically, this paper focuses on Content-Based Image Retrieval (CBIR) systems, which rely on the visual content of images to search for corresponding images in a target image database based on their features. A comprehensive literature review of CBIR approaches is presented to demonstrate the advantages of this method for managing digital images. Additionally, this paper reviews the various applications of automatic image indexing, such as photographic archive management, Digitization of audiovisual archives, e-commerce, social media, and industry. Automatic image indexing can help in organizing products by category, improving product searches, and detecting inappropriate or offensive images.