Trustworthy Authentication on Scalable Surveillance Video with Background Model Support
Zhuo Wei,Zheng Yan,Yongdong Wu,Robert Huijie Deng
DOI: https://doi.org/10.1145/2978573
IF: 4.094
2016-01-01
ACM Transactions on Multimedia Computing Communications and Applications
Abstract:H.264/SVC (Scalable Video Coding) codestreams, which consist of a single base layer and multiple enhancement layers, are designed for quality, spatial, and temporal scalabilities. They can be transmitted over networks of different bandwidths and seamlessly accessed by various terminal devices. With a huge amount of video surveillance and various devices becoming an integral part of the security infrastructure, the industry is currently starting to use the SVC standard to process digital video for surveillance applications such that clients with different network bandwidth connections and display capabilities can seamlessly access various SVC surveillance (sub)codestreams. In order to guarantee the trustworthiness and integrity of received SVC codestreams, engineers and researchers have proposed several authentication schemes to protect video data. However, existing algorithms cannot simultaneously satisfy both efficiency and robustness for SVC surveillance codestreams. Hence, in this article, a highly efficient and robust authentication scheme, named TrustSSV (Trust Scalable Surveillance Video), is proposed. Based on quality/spatial scalable characteristics of SVC codestreams, TrustSSV combines cryptographic and content-based authentication techniques to authenticate the base layer and enhancement layers, respectively. Based on temporal scalable characteristics of surveillance codestreams, TrustSSV extracts, updates, and authenticates foreground features for each access unit dynamically with background model support. Using SVC test sequences, our experimental results indicate that the scheme is able to distinguish between content-preserving and content-changing manipulations and to pinpoint tampered locations. Compared with existing schemes, the proposed scheme incurs very small computation and communication costs.