Data-Loss resilience video steganography using frame reference and data ensemble reconstruction

In this paper, we propose a robust video steganographic method, which can efficiently hide confidential messages in video sequences, and ensure that these messages are perfectly reconstructed by recipient. To apply proposed scheme to video sequences, we must be faced with two nontrivial problems: (a) how to effectively minimize the total steganographic distortion for each video frame? (b) how to recover the hidden messages if some frames are lost or damaged? We tackle the first question by designing a new distortion function, which employs two continuous adjacent frames with the same scene as side-information. The second question is addressed by data sharing. In this mechanism, the original data is expanded and split into multiple shares by using multi-ary Vandermonde matrix. Since these shares contain a lot of data redundancy, the recipient can recover the hidden data even if some frames are damaged or lost during delivery. Extensive experiments show that proposed scheme outperforms the state-of-the-arts in terms of robustness and diverse attacks.