Design of Multiuser Collusion-Free Hiding Codes with Delayed Embedding

The design of collusion-free hiding codes for a large number of users in a media distribution scenario is considered in this work. To increase the number of users, we propose to assign the same codeword repetitively to multiple users but with a different amount of delay. It is shown that, when the proposed system is under collusion, it is equivalent to a communication system with a multipath channel. Then, the identification performance for each user code can be improved by multipath channel estimation and adaptive equalization methods. We present a set of collusion-free codes called OSIFT (orthogonal spreading followed by the inverse Fourier transform) and demonstrate that OSIFT codes are robust against a generalized collusion attack in an exemplary audio watermarking system with a large number of users by computer simulation.