Abstract: “In this paper, watermark is generated uniquely for given video combining copyright information, video signature and time stamp. This watermark is scrambled using Arnold transform method and then divided into eight sub-images. Video is divided into video sequence and audio sequence. Frames are selected using scene change method, for embedding. Each scene change frame is divided into 8 x 8 blocks and DCT is applied to check energy. High energy block is selected for embedding the watermark bit. Embedding is done using Blind DCT-SVD method. Audio stream is also divided into equal size frames and DCT-SVD method used to add watermark bit to each frame. Experimental results show that proposed algorithm is useful for copyright protection as well as authentication. Embedded frame is compared with original frame to check its quality. Peak Signal to Noise Ratio (PSNR) is 49.77 dB for atrium.avi and 48.35 dB for baby.mp4. Embedded video is tested for robustness against attacks such as frame swappi”

Keywords: DCT, SVD, Video Watermarking