The first two chapters of this book have discussed the field of watermarking at a fairly nontechnical level. With the present chapter, we begin our exploration of the detailed, technical principles. In this chapter, and throughout the rest of the book, we illustrate these principles with algorithms and experiments described in special sections called Investigations.

The goal of this chapter is to familiarize the reader with several conceptual models of watermarking. These models serve as ways of thinking about actual watermarking systems. They fall into two broad groups: models based on a view of watermarking as a method of communications, and models based on geometric views of watermarking algorithms.

The chapter begins with some background material. Section 3.1 describes the basic notational conventions used in equations throughout the book. Section 3.2 gives a brief description of traditional communications systems, outside the context of watermarking. The next two sections form the bulk of the chapter. In Section 3.3, we describe three different ways of modeling watermarking in terms of traditional communications. This is followed, in Section 3.4, by a discussion of geometric models of watermarking.

Finally, we conclude the chapter with a brief discussion of correlation-based watermarking. This is a subclass of possible watermarking algorithms that we believe encompasses the majority of proposed systems. Much of the analysis presented in the rest of this book is focused on these types of systems.

A note on terminology. In the foregoing chapters, we have been using the term watermark