面向社交网络的去同步内容分发

数字指纹是一种比较完善的防止数字媒体未经授权再次分发的方法.然而,当对手收集到足够多的拷贝后,就能减弱甚至消除指纹,导致跟踪器可能无法再检测到任何实际参与合谋的人.在面向社交网络的多媒体内容分发系统中,能容纳的用户多达上千万甚至上亿,同一社区中的大量用户可聚集更多的指纹拷贝来进行共谋攻击.参与共谋的人数越多,共谋拷贝中残留的单个用户的指纹信息越少.如果不解决社交网络环境下大规模用户的共谋行为,必将造成内容所有者的隐私泄露,整个多媒体行业也会因此遭受重大损失.为了减少大多数用户参与共谋攻击后所造成的损失,提出面向社交网络的去同步指纹技术,通过在指纹媒体拷贝之间引入去同步差异性,来威慑共谋攻击行为.实验结果表明,本文提出的去同步指纹技术具有有效性.

Desynchronizated Content Distribution for Social Network

Given a large number of users in social network, the existing fingerprinting schemes decrease the efficiency of the traitor tracing and content distribution. In this paper, a new desynchronization fingerprinting for social network is proposed. The new fingerprinting technique deals with random distortions due to local geometric transformations of random grid template of image based on chaos. The experimental results showed that collusion even with only two copies resulted in disturbing dis- tortions although colluders might try to use image registration to undo the desynchronization with ex- pensive computational cost.