Authentication of Scalable Video Coding Streams Based on Topological Sort on Decoding Dependency Graph

Abstract

The security of streaming data should be ensured in current complex data era in order to provide a trusted and secure network environment. To authenticate the scalable video coding (SVC) streams by fully utilizing its decoding relationship without reducing its scalability, we establish an acyclic and directed decoding dependence graph (DDG) on the logical units of SVC steams. By applying the topological sort on DDG, we obtain the hash appendence mode for different layers of the streams (i.e., spatial and temporal layers). We propose a secure and efficient SVC authentication method based on the deduced hash appendence mode. With regard to the quality layer, we consider the corresponding quality data packets with unequal importance, and propose a grouping authentication strategy with constrained group lengths. We form the optimization problem of minimizing the authentication cost, and solve it by an iteration method. Simulation results show that our authentication approach can achieve much less computation cost and much lower overhead, while it can preserve higher verification rates and better recovered quality of video as compared with other state-of-the-art methods.