Transform Domain Watermarking Algorithm

The rapid growth of multimedia manipulation tools and network access leads to convenience of multimedia information processing, distribution and storage. At the same time, multimedia information security problems including copyrights piracy emerge as side effects of the popularity of digital representation and distribution over network. Digital watermarking has been demonstrated to be very effective in identifying the source, creator, owner, distributor, authorized consumer, or integrity of a media document (usually an image). In this dissertation, several important problems about digital watermarking are studied. Novel algorithms are proposed and the experimental results are given.The background and overview of digital watermarking are introduced in chapter 1 and chapter 2.The chapter 3 presents a few usual watermarking pre-processing methods, especially the method based on Logistic Chaos system is discussed in detail.A novel digital watermarking scheme based on DWT is proposed in chapter 4. By modifying coefficients of host image in frequency domain, a gray-scale image is embedded into host image. The approximate sub-band of host image is divided into many blocks and some of these blocks are selected with HVS. Watermark is embedded into host image by increasing a coefficient and reducing another coefficient in selected blocks, by this way, watermark can be extracted without initial host image. Experimental results show that proposed algorithm provides good invisibility and robustness.A practical application about how to realize watermarking system that is proposed in chapter 4, which is based on DSP6000, is proposed in chapter 5. In the processing, programs make full use of the advantages of DSP hardware resource with considering practical application of watermarking system. Watermarking system is divided into two parts: embedding and extracting. Each part includes some common programs and especial programs, the common programs such as reading and saving image, DWT and IDWT and some others. Especial part only includes embedding and extracting. By this way, the system is flexible. The system can accomplish watermark embedding and extracting in real time. It makes sense to apply watermarking technique in practice.Finally, the prospect of watermarking technique is depicted in chapter 6.