An optimized un-compressed video watermarking scheme based on SVD and DWT

In this paper, we presents a novel fast and robust video watermarking scheme for RGB uncompressed AVI video sequence in discrete wavelet transform (DWT) domain using singular value decomposition (SVD). For embedding scene change detection is performed. The singular values of a binary watermark are embedded within the singular values of the LL3 sub-band coefficients of the video frames. The resultant signed video exhibits good quality. To test the robustness of the proposed algorithm six different video processing operations are performed. The high computed PSNR values indicate that the visual quality of the signed and attacked video is good. The low bit error rate and high normalized cross correlation values indicate a high correlation between the extracted and embedded watermark. Time complexity analysis shows that the proposed scheme is suitable for real time application. It is concluded that the embedding and extraction of the proposed algorithm are well optimized. The algorithm is robust and shows an improvement over other similar reported methods.