A NOVEL DIGITAL VIDEO WATERMARKING TECHNIQUE BASED ON WAVELET, SVD AND CZ-TRANSFORM

Abstract

Digital Watermarking is a technology of embedding data into digital multimedia content to verify the credibility of the content or to recognize the identity of digital content owner. To identify the ownership of the digital content a new digital video watermarking algorithm is proposed. The proposed algorithm divides frames of a cover video into three colour bands of red, green and blue. Then the accompanying three tasks are performed on every one of three colour bands independently. Initially, each colour band is divided into blocks of small sizes and afterward the entropy of each block is computed. At this step, a threshold is found based on the average entropy of all blocks and following is applied to all blocks associated with lower entropy than the threshold. A wavelet representation of each block is given by applying a discrete wavelet transform. Then Singular value decomposition, orthogonal-triangular decomposition, and a chirp z-transform are used to embed a watermark on the cover video. To check the robustness of the algorithm different signal processing attacks are applied on watermarked video. The Proposed algorithm is compared with singular value decomposition based blind watermarking algorithm. Experimental results shows that proposed algorithm outperforms the singular value decomposition based blind watermarking algorithm quantitatively in term of PSNR.