Study On Image Watermarking Algorithms Based On DCT And Fractal Coding

With the development of computer networks and the communication technology, digital works,such as image, video and audio,may be obtained easily, at the same time care is taken to protect copyrights of these works and prevent the pirate. Therefore, digital watermark technique is brought forward to protect the legitimate right of owner of work. Fractal image coding is a very promising image compression technique. Due to its novelty, high compression rate, resolution independence and fast decoding, the technique has received more interests from the researchers in the area of image processing, especially image compression over the past decade. Now, fractal image coding has spread over other areas of image processing.In this dissertation, some basic knowledge of digital watermarking techniques is first systematically discussed, including its background, general system structure, basic principle, state-of-the-art, and their applications. Two adaptive image watermarking algorithms combining Discrete Cosine Transform (DCT) and fractal coding are then proposed. In detail, the host image is split into non-overlapping fixed-size square blocks, which are categorized into two classes according to their collage errors obtained by performing baseline fractal coding. The blocks classification is utilized to form a feature mark of the host image and determine imbedding strengths of the watermark in distinct blocks, and Arnold transform is applied to scramble both the feature mark and binary copyright symbol. The watermark is the fusion of the scrambled feature mark and a scrambled binary copyright symbol (using XOR operation). Last, the watermarks are embedded into the middle frequency bands of DCT domains of distinct blocks by comparing the selected three coefficients or qualifying the DC coefficients.In order to test the imperceptibility and robustness of the algorithms, computer simulations on the two algorithms are conducted respectively. The simulation results indicate that the watermarking algorithms are both imperceptible and robust to common attacks such as JPEG, cropping, Gaussian filtering, noising, etc.