Improving a Digital Rights Management Scheme for Video Broadcast

The security of content distribution is a critical issue that all the content providers have to treat seriously in order to guarantee their business. Emmanuel and Kankanhalli proposed an efficient videobroadcast scheme in [1] which is able to distribute video to many subscribers by multicasting one video copy only. The scheme is also designed to achieve the security such that the illegal distributors can be identified based on the client-side watermarking. In this paper, we present an attack which enables any eavesdropper to create a pirate video of good quality from the protected video. Concretely, an adversary can estimate the secret interference watermark which is used to protect the distributed video by exploiting the statistics of randomly selected video frames/images. To fix the flaw, we propose two alternative countermeasures. One is to encrypt the protected video in transmission, and the other is to modify the interference signal frame by frame. Both countermeasures can disable an eavesdropper to obtain a video of acceptable quality, but maintain the nice features that the original scheme is designed for, i.e., efficiently multicasting the protected video with only one channel; managing the broadcaster copyright; and preventing from framing innocent subscribers.