Digital Watermarking of video streams: Review of the State-Of-The-Art (1908.02039v2)

Abstract: Digital Watermarking is an extremely wide aspect of information security, either by its applications, by its properties, or by its designs. In particular, a lot of research has been made about video watermarking and it can make it quite difficult to put into perspective the various schemes possible in order to implement a watermarking process for a given application. This paper presents an in-depth overview of the current video watermarking technologies and how they each respond to certain criteria that may be imposed by the aimed application. The goal being in first place to be able to define the desired equilibrium point between invisibility, robustness and efficiency for an application. Then, given this balance, being able to deduce the best location of the information embedding as well as the method used to embed it. The equilibrium point is to be found using the needed properties of the watermark and by studying the threat model that the scheme will have to face. The location describes whether the extra information should be added to the metadata of the video, to its frames or to specific regions of its frames. Finally, the method to embed the watermark refers to the insertion domain and its coefficients to be altered in order to insert the wanted information.