Perbandingan teknik watermarking berbasis SVD, DCT-SVD, dan SVD-DCT pada video digital

Abstract

With the advantages of digital video processing, digital video can be manipulated and distributed illegally by irresponsible parties. Such actions are detrimental to owners of digital video and violate the copyright of the digital video. One solution to overcome this problem is to apply the watermarking technique in digital video. Watermarking techniques can be performed in spatial domain, frequency domain and hybrid domain. In the spatial domain, one of the watermarking techniques can be performed using Singular Value Decomposition (SVD) as proposed by Liu-Tan. In the frequency domain, one of the watermarking techniques can be performed using Discrete Cosine Transform (DCT). In the hybrid domain, one of the watermarking technique can be done by combining SVD with DCT. In this paper, the author tries to apply the watermarking technique proposed by Liu-Tan on digital video in spatial domain and hybrid domain. There are two methods in the domain of hybrid that will be proposed, namely the transformation of DCT-SVD and SVD-DCT transformation. Three watermarking technique compared based on the PSNR value of the insertion, insertion time, the correlation value from extraction, extraction time, and the robustness. Attacks carried out in this study is the attack noise, filtering attack, attack rotations, and cropping attacks. The results showed that the watermarking technique proposed by Liu-Tan is applied to the hybrid domain would increase the value of PSNR and correlation values, but vulnerable to some attacks. In the hybrid domain, the transformation of SVD-DCT is better than DCT-SVD transformation.