Performance study and improvement on ECC-based binary anti-collusion forensic code for multimedia

Digital forensic coding is an emerging technology that offers proactive post-delivery protection of multimedia. Multi-user collusion attacks are powerful attacks against digital forensic marking, where groups of attackers collectively mount attacks to attenuate the identifying marks. The colluders usually perform post-processing, such as compression, on the colluded multimedia content before redistribution. The post-processing may introduce errors in the detected bits of the forensic code and represents a new challenge to the forensic code design. In this paper, we first study the performance of a simple extension from ECC-based forensic marking to create a binary forensic code, and then improve its performance to provide post-processing resistance. This new code is constructed by concatenating an orthogonal binary code with a large-distance Reed-Solomon code. The simulation results show that our binary forensic code design can achieve 100% detection rate when the number of colluders is less than 20 and up to 32% of the colluded forensic code is changed during the post-processing.