On the capacity game of private fingerprinting systems under collusion attacks

The problem of fingerprinting (FP) in the presence of collusive attacks is considered. It is modelled as a game between a fingerprinter and a decoder, on the one hand, and a coalition of two or more attackers whose action is modelled by a multiple access channel (MAC), on the other hand. A single-letter expression for the capacity is derived under mild conditions, asymptotically optimum strategies of the parties involved are characterized, and a lower bound on the error exponent is obtained.