The rapid growth of the Internet as a distribution medium for digital goods increased the risk of copyright infringements. From an economic point of view, this risk makes the commercialization of digital works difficult, if not impossible. Therefore, the need for technical copyright protection solutions has increased steadily over the last years. Robust digital watermarking became a promising technology in the context of copyright protection and was proposed as a central building block in various e-Commerce protocols (such as dispute-resolving, copy protection and traitor tracing schemes or DRM applications). Traditionally, the design of watermarking schemes was seen as a signal-processing problem and concentrated on issues such as the imperceptibility of the watermark or its resistance against unauthorized removal. However, when a watermark is to be used in an e-Commerce system, its properties may become critical to the security of the overall scheme. It is therefore necessary to gain a thorough and mathematically precise understanding of the essential security properties of watermarks.
In this talk, I review recent results [1,2] that establish the security of two cryptographic protocols that employ watermarking operations as basic primitives. The first protocol can be used in dispute-resolving schemes in order to assure their resistance against an important class of attacks. The second protocol allows to detect forgeries in image files or video streams by embedding a watermark carrying a cryptographic signature. Both constructions are provably secure under standard cryptographic assumptions.