Automating the linking of content and concept

Abstract

In previous work we have described a multimedia system, MAVIS 2, supporting content and concept based retrieval and navigation. A central component of the system is a multimedia thesaurus in which media content is associated with appropriate concepts in a semantic layer. A major challenge is identifying and constructing these associations in a particular application without requiring a huge amount of manual effort. In this paper we propose a two phase approach to the problem. In the first phase, latent semantic analysis is used to associate metadata available for some media objects with concept class descriptions. This facilitate automatic associations to be made with the concept layer of the those media objects. In the second phase, media content matching is used to classify media objects without metadata through their similarity to media objects classified in phase 1.