Please use this identifier to cite or link to this item: http://hdl.handle.net/1783.1/7335

Perceptual video encryption system : a multiple transforms approach

Authors Au Yeung, Siu Kei
Issue Date 2011
Summary With the increase of multimedia applications on the internets, security in digital video information has become more important. For instance, in commercial Video-On-Demand (VoD), PayTV, and Pay-Per View applications, it is very critical that only the customers who have paid for the service can enjoy the high-quality video. Since video data are transmitted through public internet, it is thus possible that one un-authorized person somehow obtains the data during the transmission. On the other hand, it would also be a good idea to provide a down-grade version of videos to potential customers as preview before they decide whether they want to subscribe to the service or not. As such, the idea of Perceptual Video Encryption has been introduced in the past few years. Different from traditional fully confidential video encryption, perceptual video encryption does not require the encrypted video signal to be robust to all the complicated and time consuming security attacks. The main aim is to introduce a fast and efficient algorithm which provides a certain level of protection to the video signal. In addition, as the encrypted signal will also be used as the preview, quality assessment of this downgrade video signal is also another important issue to consider. In this thesis, we propose that the perceptual encryption can be performed at a stage within the video encoding process that had never been considered before: the transformation stage. To this end, we first design a number of new unitary transforms based on some plane-based rotations in the flow graph structure of the well-known discrete cosine transform (DCT). We demonstrate that these new transforms are performing as efficiently as the DCT does in the coding of predicted residual signals. Then, our new perceptual encryption is constructed by using these new transforms together with the DCT but alternately, i.e., we select one out of multiple transforms in each video block according to the encryption key. This algorithm is tested extensively with the H.264 and MPEG-4 codec. Both the PSNR-based and SSIM-based results are presented to demonstrate the effectiveness of our new encryption algorithm. In order to evaluate the proposed scheme fairly, a full analysis of various assessment schemes is performed. In addition, several improved extensions on the proposed scheme are presented. Finally, security analysis is also performed to test the resistance of our algorithm to different types of attacks.
Note Thesis (Ph.D.)--Hong Kong University of Science and Technology, 2011
Subjects
Language English
Format Thesis
Access
Files in this item:
File Description Size Format
th_redirect.html 337 B HTML