Enhancing the robustness of digital video watermarking for copyright protection against collusion attack

Abstract

In the digital world, due to the ease of creating, publishing and tampering of works, copyright protection became a major issue. This issue is much significant in movie industry which cause millions of dollars loses. With new technologies almost everybody are able to produce, edit and share videos. Thanks to the Internet the sharing of works also, become a breeze. Hence, there should be a mechanism to prove the ownership and make sure of the integrity. Watermarking is one of the effective ways to help copyright protection. According to this problem, this project is defined to enhance the current copyright protection. We narrow down the project to a specific attack. Collusion attack is the main problem for video watermarking for copyright protection. In this attack, colluders try to estimate and remove the watermark by comparing as many as watermarked materials. By increasing in number of watermarked materials, the estimation becomes more accurate. The proposed method is a frame-by-frame video watermarking. To address the collusion attack, each frame receives a different watermark. Since each frame has a unique watermark, it is impossible for the attack to estimate it. In this project Block Truncation Coding (BTC), Discrete wavelet transform (DWT) and Singular value decomposition (SVD) are combined to achieve the required robustness against the attack as well as maintaining quality. The BTC is used in watermark generation. DWT and SVD are used to prepare the frames and embed the watermark. The robustness is tested using Bit Error Rate (BER) and the quality is checked using Peak signal-to-noise ratio (PSNR). The results of the project, in compare with related works, shows that the proposed method is robust against all types of collusion attack, and at the same time, the quality of the video is remain in a satisfactory level. Meanwhile, it is proved that the method is robust against some other attacks as well.