Research On Image Watermarking Against Geometric Attacks Based On Dual Transform Domains And Fractal

Digital watermarking is a potent method for multimedia content protection such as protection of copyright, and becomes significant disquisitive field in the information hiding.But,at present,not every watermarking algorithm is rubust to all the attacks.The aim to designing watermarking algorithm is to against a particular type of attacks,and try to take other types of attacks into account. As the watermarking product application areas are often varying,and different watermarking application environments tend to a certain algorithm against a type of attacks.Therefore,the research for a certain type of attacks watermarking algorithm is also of critical importance.This article is focusing on cutting,removal of rows and columns and rotatation geometric attack, three kinds of algorithms were designed.1.A meaningful digital watermarking algorithm is proposed based on combination of DCT and DWT.Firstly,making watermarking chaos using arnold transform;Secondly, cutting the choas into two pieces. Second, embeding one piece into DCT's DC coefficie-nt coefficient by adaptive quantization,and embeding the other into DWT sub-band.2. A new fractal watermarking algorithm robust to removal is proposed based on image normalization and fractal.Firstly,the geometrically invariant space is built by using image normalization.Then,a significant region is gained from the normalized image by using invariant centroid principle.Finally, the watermark is embedded into the self-similar blocks DCT transformed of the significant region by fractal.3. A new image watermarking algorithm robust to different angle rotation is proposed.Firstly,significant region is extracted by using normalization and invariant centroid principle.Then,significant region is divided into two groups:self-similarity and non-self-similarity blocks by fractal coding and pre-set threshold,the robustest pseudo-Zernike moments is picked up among the pseudo-Zernike moments of the self-similarity blocks.Finally,the watermark is embedded in the robustest pseudo-Zernike moments by quantization.