A Robust and Secure Signature Scheme for Video Authentication

In this paper, we present a hash-based technique for video authentication. Designing a good hashing technique for multimedia requires many issues to be considered such as robustness, security and tamper detection with precise localization. The work that we shall describe in this paper is a step ahead in this direction that attempts to simultaneously address these issues. The hash of the video is generated using the features extracted in the DCT domain. These features are then quantized such that the quantized features remain same for a bounded distortion. To enforce security, pixels of the luminance plane are directly modulated using a secret key. We present experimental results to demonstrate the robustness, fragility and security of our proposed scheme.