Game-Theoretic Analysis of Maximum-Payoff Multiuser Collusion
Multiuser collusion is an effective attack against traitor-tracing multimedia fingerprinting, where a group of attackers collectively mount attacks to reduce their risk of being detected. During collusion, each attacker wishes to maximize his or her own payoff. To resolve the conflict, colluders have to negotiate with each other and achieve fair collusion. An attacker also needs to decide with whom he or she wants to collude. Though colluding with more people helps further reduce the risk, it also makes an attacker share with more people the rewards from illegal usage of multimedia. This paper uses game theory to model the complex colluder dynamics and analyzes the tradeoff between the risk and the rewards.