Efficient multi-layer coding and encryption of MPEG video streams

This paper introduces a novel, light-weight video encryption algorithm that supports light-weight, multi-layered encryption. The objectives of this encryption algorithm are to reduce the total amount of data encrypted (while providing reasonable privacy and security) and to allow for the playback of the encrypted stream in the presence of network packet loss and bit-errors. The latter property allows for the easy adaptation of encrypted video over best-effort networks, such as the Internet. This algorithm partitions the stream into three layers and provides encryption on the lower two layers. An adaptive algorithm is provided that shows how to adaptively partition the video data so that the user can ensure a maximum peak signal to noise ratio in the base layer. Our results show that we can provide security by encrypting only a fraction of the data depending on the level of security the user requires.