Overview

The security of a watermark refers to its ability to resist intentional tampering, as opposed to the common signal distortions discussed in Chapter 8. Intentional tampering or hostile attack becomes an issue when someone is motivated to prevent a watermark from serving its intended purpose. Watermarks can be attacked in a variety of different ways, and each application requires its own type of security.

We begin, in Section 9.1, with a discussion of the types and levels of security that might be required of a watermarking application. We then discuss the relationship between watermarking and cryptography, in Section 9.2. We find that there are cryptographic tools that can be directly applied to watermarking systems to achieve some types of security. Those security areas in which cryptographic tools are helpful are distinguished from those that must be addressed in other ways. Finally, in Section 9.3, we describe a number of well-known attacks on watermarking systems. For some of these attacks, reliable countermeasures are known, but security against some others remains a topic of research.

9.1 Security Requirements

The security requirements for a watermarking system vary greatly from application to application. In some applications watermarks need not be secure at all, because there is no motivation to tamper with or circumvent the watermark's intended pupose. For example, most of the device-control applications described in Section 2.1.7 use watermarks only to add value to the content in which they are embedded. In such cases, no one benefits from tampering, and...