Indexing Multimedia and Creative Works: The Problems of Meaning and Interpretation

In principle, any object that exists (or has existed) – be it transient or long‐lived, static or dynamic, abstract or physical – can be indexed and details of it (or the artefact itself) then subsequently retrieved from some sort of archival storage system. The storage system that is used might be a collection of cardboard boxes, a museum, an art gallery, a library or a computer system. Of course, for transiently existing objects, or those that have long ceased to exist, it may be that only representations of the artefact of interest (and/or commentary on it) can be stored and retrieved. Despite its conceptual simplicity, object/information retrieval is by no means a trivial issue. This is especially true when one considers the difficulties of cataloguing and retrieving multimedia and creative works – such as novels, plays, music and film material. Within this book, the authors outline some of the problems involved in indexing multimedia materials and subsequently retrieving details of them from archival storage.

The book contains six chapters and an extensive bibliography/reference list. The first two chapters are essentially introductory. The first chapter is used to introduce the overall theme with which the book deals; it discusses the concepts of information (and its retrieval) within the context of discourse and communication. The second chapter then goes on to provide an overview of information retrieval (IR) from two basic perspectives: the theoretical stance taken by information scientists and librarians; and the more pragmatic approach based on the use of computers. This chapter identifies some of the key themes from text‐based IR (such as classification, indexing, cataloguing, Dublin Core, and so on) that might be usefully applied to multimedia information retrieval (MMIR). This latter topic is discussed in considerable depth in chapter 3. For many different reasons, MMIR is much more difficult to implement than text‐based IR. One of the main reasons for this is the fact that there is no single, universal multimedia indexing standard in common use. In this chapter, several different approaches are discussed – particular emphasis being given to the importance of markup techniques (SGML and XML).

In Chapter 4, the authors discuss the use of semiotics for analysing multimedia objects. This is an interesting chapter which illustrates how semiotic analysis can be used to derive meaning from image‐based objects and sound artefacts. Chapter 5 is by the largest one in the book. It describes a number of multimedia retrieval tools and illustrates their use for indexing some examples of multimedia objects. The authors also use semiotic analysis in order to illustrate the range of meanings that might be associated with multimedia objects. The final chapter in the book deals with research issues in MMIR. Two broad fronts are covered: content‐based image retrieval; and concept‐based retrieval involving “democratic indexing” involving a pictorial semiotics framework. The authors describe the latter as: “a philosophical approach to MMIR which attempts to retain some of the richness of free description and association which semiotic analysis permits us and controlled vocabulary curtails”.

This is an interesting book that “attacks” a difficult and important problem ‐ the intractability of which is likely to increase significantly in coming years as we all face the mounting issues of retrieving our personal collections of static images, cuttings, multimedia creations, DVDs and sound recordings.