Tagging of Multimedia Contents on the Web 3.0 Using Semantic Artificial Intelligence: A Systematic Literature Review

With the World Wide Web transitioning from Web 2.0 to the more intelligent and reliable Web 3.0, the importance of effective tagging methods for multimedia data has become increasingly important. In contrast to text documents, where keyword extraction is quite straightforward, tagging multimedia objects such as images, audio, video, and infographics is computationally expensive and highly demanding. This paper thoroughly defines the common multimedia data tagging methods and contrasts their effectiveness in semantic intelligence systems as well as usage in general and domain-specific scenarios. Systematic searching, inclusion, and exclusion were used to choose 90 peer-reviewed articles from an initial pool of 136 papers. The paper then contrasts the performance of current multimedia tagging technologies based on 11 benchmark datasets. This review paper concentrates specifically on semantic-focused artificial intelligence (AI) and factual knowledge as major enablers towards enhancing the scalability and accuracy of multimedia tagging in the highly coherent Web 3.0 environment. The survey points out the implications of knowledge graphs (KG), ontologies, and semantic reasoning (SR) in boosting semantic comprehension, as well as outlines how these technologies tackle issues such as serendipity, overspecialization, and the cold start problem. Through the synthesis of insights from current work, this paper points out gaps in prevailing strategies and outlines paths towards the construction of computationally feasible and trustworthy semantic models for multimedia tagging in the future. The survey seeks to offer a general perspective on the state-of-the-art in multimedia tagging and contribute to the development of the Semantic Web as a web of interconnected and contextually informative data.