This timely book provides surveys in important areas related to multimedia information retrieval systems (MIRS). The book’s audience includes graduate students, researchers, and practitioners who are interested in MIRS.

The book consists of 13 well-organized chapters. The contents of the chapters are somewhat independent. Blanken, de Vries, Blok, and Feng make a good effort to connect the chapters, by stating the relation of each chapter to the rest of the book, which makes it easier to read.

The book starts with an introduction to the common multimedia we see on the Internet (audio, video, and still images) and an overview of general MIRS. The introduction chapter also states the theme of the book: how to index and retrieve multimedia data from an information retrieval point of view.

In the next four chapters, Blanken et al. present a set of basic building blocks of any MIRS system. One important aspect of MIRS is the ability to index and retrieve information through the use of metadata--typically plain text that describes the multimedia content. Chapter 2 deals with languages for metadata. Chapter 3 introduces pattern recognition as a tool for multimedia content analysis. Chapter 4 discusses the issue of text-based index and search, and chapter 5 concentrates on one type of media: images.

Chapter 6 presents the generative probabilistic model that can be used to index and retrieve any type of multimedia content. The following two chapters discuss audio and video indexing. Two different content-based video retrieval models (spatio-temporal and probabilistic approach) are the subject of chapters 9 and 10.

Chapter 11 introduces readers to four different types of interactions between a typical user and a MIRS: retrieval, dynamic query interaction, browsing, and recommendation. It also presents interactions from other perspectives, such as relevance feedback, personalization, and content adaptation.

The book closes with two more relevant chapters, one dealing with digital rights management and the other handling performance evaluation of MIRS.

The book is a very good reference for both researchers and practitioners. If used as a textbook, students should consult supplementary materials, such as papers or other textbooks, as some of the necessary information is not sufficiently detailed. It would be more complete if the authors discussed some popular MIRS systems in greater detail.

Overall, this is a very good book and I enjoyed reading it. It covers a complete list of topics relevant to MIRS. The further reading sections at the end of each chapter provide useful information to explore the topics in more depth. The references are current and abundant.