Study On Gray-scale Image Watermarking Algorithms And Its Theoretical Analysis

With the development of digital multimedia technology, digial watermarking has become a popular topic in the field of digital copyright protection. Digital watermark protects the securities of digital multimedia by embedding special information into it. Applications of watermarking include not only copyright protection, but also document authentication, covert communications, data embedding and so on. In these applications, a watermark is embedded into a host data. This embedding should be nearly imperceptible and robustness against possible manipulations of the watermarked data, in the sense that it should be possible to reliably extract the watermark from a degraded version of the signal. Degradations include operations such as addition of noise, filtering, compression and geometrics distortion. These degradations could be intentional or nonintentional. The designing of a watermark resistant to these distortions is still an open problem. Robustness is crucial for watermarking technology in order to be used widely in commercial areas.In this dissertation,we take still gray-scale images to study the technology of digital watermarking as examples. The key points of research addressed in this dissertation are as following:Using modern communication technology (multiple input multiple output thought) for reference, we design a robust gray-scale waternark algorithm. The key of designing this algorithm is how to effectively select the image's characteristic and embedding formula to minimize the damage to original image when the gray-scale watermark is embedded many times, and how to effectively use the two extracted watermarks to get the best quality watermark. To tackle these problems, we proposed a wavelet-based watermarking technique, which is combined with fractal coding. First, a host image is partitioned into a set of non-overlapping square blocks. These blocks are sorted to become a descending sequence according to their collage errors which are obtained by performing baseline fractal coding, and two equal length subsequences are respectively chosen according to the watermark size. Second, the watermark, which is first scrambled using Arnold transform, is embedded into low-frequency bands in wavelet domains of distinct blocks in the two selected subsequences by a new embedding formula. Last, the watermark is the fusion of the extracted watermarks from the two selected subsequences. Besides, we also analyze theoretically the proposed embedding formula and the second fusion method. The experimental results show that this technique achieves good perceptual invisibility and security, and is also very robust against some image processing such as cropping, addition of noise, scribbling, filtering, rotation and JPEG compression.To a kind of gray-scale watermarking system, we analysis the relationship of the embedding strength, the robustness and the invisibility, from which we obtain the quantitative descriptions among them and give a compromise between invisibility and the robustness. These results provide certain theory instruction for the design of watermarking algorithms.