Compressed Video Watermarking for Authentication and Reconstruction of the Audio Part

This paper is concerned with designing a digital video watermarking system capable of authenticating and reconstructing the audio part of the video after possible attacks. As recompression is the most common attack on videos, we attempt to improve robustness of one of the recently presented and successful compressed video water-marking schemes against recompression. A comprehensive set of experiments are conducted to show that our watermarking scheme is robust against recompression attack and enables reconstruction of audio part with an acceptable quality based on PESQ score. To the best of our knowledge, this is the first work that uses the visual part of a video as a watermarking cover signal for tampering detection and reconstruction of the audio part. We specifically address videos with important audio content, e.g., news, reports, etc.