Study On Digital Watermark Of Images Based On DWT

Intellectual property protection receives much concern both at home and abroad, and some relative problems are difficult to solve, especially the protection of digital media copyright. These digital media including images, texts, audio and video digitalized works are easy to store, design, transmit and duplicate, and this accelerate communication and transmission of knowledge and information. But they also offer convenience to some illegal persons who can distort, attack and pirate the achievement maliciously at the same time. With the rapid development of internet, multimedia and communication technology, protection of digital media copyright has been paid more attention than ever before, which makes the digital watermarking technology concerned by people in particular. As a novel technology, digital watermark rises in recent years, and offered brand-new thought to realize effective protection of intellectual property. Now it has become a research hotspot in the field of multimedia information security.This dissertation focuses on wavelets and their application algorithms for image digital watermarking.In DWT domain, most energy of image converge in approximation subband, when large amounts of data are embedded in this subband, both the PSNR value and the visual quality of the watermarked image drop greatly. So we embed watermark in detail subbands. Though many DWT-based watermarking algorithms are presented now, few of them discuss the issue that which detail subband is a better choice to embed watermark. Through a large number of experiment, we draw a conclusion that watermarking in the subbands with higher RMS value is more robust to Gaussian noise and JPEG compression. For the sake of further robustness of the algorithm, we choose some stable coefficients in the subbands to watermark. These selected coefficients change little after signal processing. This algorithm was implemented by MATLAB 6.5. The experimental results show that this algorithm isinvisible and robust against noise and commonly image processing. So it has practical value.