Cryptographic methods for collusion-secure fingerprinting of digital data

We give a survey of cryptographic techniques for collusion-secure fingerprinting. Fingerprinting of digital data is a method for copyright protection in which an individual version of the data to be sold is created for each buyer by the introduction of minor errors, called marks, into the data. If the method guarantees that even collusions of dishonest buyers are not able to create versions of the data which are untraceable it is called collusion-secure fingerprinting. D. Boneh and J. Shaw gave [Proc. of CRYPTO'95, Springer, Berlin, 1995, pp. 425-465] a model for collusion-secure symmetric fingerprinting and the most efficient method known so far for its realization. B. Pfitzmann and M. Schunter extended this model [Proc. of EUROCRYPT'96, Springer, Berlin, 1996, pp. 84-95] by the invention of asymmetric fingerprinting which protects innocent buyers from accusation by cheating merchants. We will explain techniques presented in Ref. I. Biehl and B. Meyer [Proc. of STACS'97, Springer, Berlin, 1997, pp. 399- 412] which allow to use each arbitrary symmetric fingerprinting scheme for the construction of an asymmetric fingerprinting scheme of almost the same efficiency (measured in the number of watermarks included in the marked data). Moreover, we explain the idea of anonymous fingerprinting presented by B. Pfitzmann and M. Waidner [Proc. of EUROCRYPT'97, Springer, Berlin, 1997, pp. 88-102] which achieves collusion-secure fingerprinting of data while protecting the anonymity of the buyers. Finally, we sketch a construction for anonymous fingerprinting schemes.