Shot-Based Content Adaptive Robust Digital Video Watermarking Algorithm Against Camera Recording
Jiayi Bai, Xin Wang, Yana Zhang, Cheng Zhong Yang · 2022
Digital video is one of the most important media in current communication. It is very popular for its rich content and vivid forms. With a rapid development of the Internet, the progress of computer technology makes communication more convenient, and also facilitated infringers' illegal and pirated acts. The copyright protection of digital video has become an urgent problem to be solved. Camera recording is a serious threat to the security of video content. The camera recording is not a single attack, but a combined attack in the spatio-temporal domain. It is difficult for a video watermarking algorithm to be robust to combined attacks in the spatio-temporal domain and satisfies the content adaptation. In this paper, we proposed a video watermarking algorithm based on the temporal domain, which takes the video shot as an embedding unit. We introduce a new metric called chrominance standard deviation (ST) to represent the changes in chrominance amplitude within the shots. According to the chrominance characteristics of the video, the embedding strength and period are determined adaptively, and the watermark is embedded in every frame of a shot. In the watermark extraction, we introduce geometric correction and color correction modules to correct the distorted video. The experimental results show that the proposed algorithm has high robustness not only to various single common attacks but also to mobile phone camera recording, while satisfying the imperceptibility of the watermark.