|
|
|
|
The ubiquity of broadband networks and the advance in multimedia technologies have inspired people all over the world to enjoy and share multimedia over networks. However, the same tools that enable us to create and distribute multimedia easily also facilitate the illegal alteration, repackaging and unauthorized redistribution of multimedia. These illicit copies pose serious threats to both governmental operations and commercial applications. Cryptographic tools and access control support the secure delivery of multimedia over networks, whose protection usually terminates after the content is delivered and decrypted. To address the post-delivery protection of multimedia, digital fingerprinting is an emerging technology to identify users who have legitimate access to the plaintext content but use it for unintended purposes. It provides a proactive way to trace the illegal usage of multimedia by seamlessly embedding unique identification information (“fingerprint”) into the content before distribution. Ensuring the appropriate use of multimedia content, however, is no longer a security issue with a single adversary. A group of attackers with differently fingerprinted versions of the same content can collectively mount attacks (known as multi-user collusion attacks) and effectively remove the traces of the fingerprints. Multimedia fingerprints should not only be robust against attacks by a single attacker, they should also resist such multi-user collusion. This requires the digital rights enforcer to have a profound understanding of multi-user collusion and design anti-collusion fingerprints for multimedia.
Figure. The general frame of embedded fingerprinting for multimedia. We take an interdisciplinary approach to understand the challenges and analyze the performance of the digital fingerprinting technology for multimedia content protection. Our research addresses the following issues in multimedia fingerprinting and traitor tracing: Multi-user collusion attacks: linear versus nonlinear collusion There are several types of collusion attacks that may be used against multimedia fingerprints. One method is simply to synchronize the media signals and average them, which is an example of the linear collusion attack. Another collusion attack, referred to as the copy-and-paste attack, involves users cutting out portions of each of their media signals and pasting them together to form a new signal. Other attacks may employ nonlinear operations, such as taking the maximum or median of the values of corresponding components of individual copies. We examine various types of collusion, including the linear averaging and a few nonlinear collusion attacks, and analyze their effectiveness in removing the fingerprints and the perceptual quality of the colluded copy under different collusion attacks. Traitor tracing capability of multimedia forensic systems In order to facilitate the design of multimedia forensic systems for applications with different protection requirements, one critical research direction is evaluating the resistance performance of specific fingerprinting schemes when considering different types of attacks. We evaluate the fundamental limits of orthogonal multimedia fingerprints, and evaluates the maximum number of colluders that digital fingerprinting systems can withstand as a function of the system parameters, including the fingerprint length, the total number of users, and the system performance requirements. Such analysis enables the digital rights enforcer to understand which factors limits the traitor tracing capability of the forensic systems and provides important guidelines for collusion resistant multimedia fingerprint design.
Secure fingerprint
multicast
The popularity of networked multimedia systems
relies on the ability to provide reliable, low-cost, and secure services to all
parties in the systems. For networked video applications where a huge amount of
data have to be distributed to a large number of users, bandwidth efficiency is
an important issue to be addressed by the service providers. In order to
accommodate more users, multicast is often used to minimize the communication
cost in the data distribution. For streaming applications with traitor tracing
requirement, employing digital fingerprinting technology poses new challenges to
the secure and efficient distribution of multimedia. This comes from the fact
that traditional multicast technology cannot be directly applied to
fingerprinted multimedia since copies distributed to different users contain
unique fingerprints and are not identical. Our work collectively addresses the
security and bandwidth efficiency for networked multimedia systems, and
investigates the secure fingerprint multicast technology.
|
Successful "traitor tracing" in real life: