As introduced in Chapter 2, multimedia fingerprinting systems involve many users with different objectives, and users influence one another's performance. To demonstrate how to analyze human behavior, we take the equal-risk fairness collusion as an example.

During collusion, colluders contribute their fingerprinted copies and collectively generate the colluded copy. As demonstrated in Section 2.3.2, depending on the way attackers collude, different colluders may have different probabilities to be detected by the digital rights enforcer. Each colluder prefers the collusion strategy that favors him or her the most, and they must agree on risk distribution before collusion. A straightforward solution is to let all colluders have the same probability of being detected, which we call equal-risk fairness. Depending on the fingerprinting system and the multimedia content, achieving equal-risk collusion may be complicated, especially when colluders receive fingerprinted copies of different resolutions, as shown in Section 2.3.2. In this chapter, we use equal-risk fairness as an example, analyze how colluders select the collusion parameters to achieve fairness in scalable video fingerprinting systems, and provide a simple example on behavior dynamics analysis and the investigation of the impact of human factors on multimedia systems.

In this chapter, we investigate the ways in which colluders distribute the risk evenly among themselves and achieve fairness of collusion when they receive copies of different resolutions as a result of network and device heterogeneity. We also analyze the effectiveness of such fair collusion in defeating the fingerprinting systems.