| As the digital form of multimedia data is developing rapidly, the access and transmission of private data become more and more convenient. However, the multimedia information can be arbitrarily edited, illegally spread and copied, and it is hard to be controlled effectively. The research significance of digital watermarking technology becomes more and more clear. The research achievements provide support for authentication and anti-counterfeit of the multimedia data. Therefore, it is of great realistic meaning and practical utility. This paper focuses on the digital watermarking algorithm in wavelet transform domain. The focus of our work is on the pre-processing of the watermark, the inserting of the watermark and extracting of the watermark. On the premise of invisibility, combining with the correlative technology, our work focuses on the robust digital watermarking algorithm in wavelet transform domain with the goal of improving robustness, and performing the simulation experiments as well. The primary contents of this paper are as follows:First, the development process of digital watermarking technology is introduced from a macroscopic view in this paper. Based on this, the two types of digital watermarking algorithm are compared, and the advantages and disadvantages of these methods are summarized. Then by analyzing the advantages of discrete wavelet transform applied to the watermark, digital watermarking algorithm in wavelet transform domain is confirmed to be focused in this paper.Second, the concept of digital watermarking, characteristics, application domain, basic framework and pre-processing are introduced in this paper. The attack methods and evaluation standards of the image watermark are also elaborated. In addition, the theory of wavelet transform and two-dimensional discrete wavelet transform are expounded, and take Lena image for example, the energy distribution of components through wavelet decomposition are proved by experiments. Some reasons were given for the digital watermarking algorithm based on low frequency components which is introduced in chapter three.Third, in order to improve algorithms that the watermark is embedded into DWT coefficients directly, a kind of wavelet domain digital watermarking algorithm based on threshold classification and fruit fly optimization algorithm is proposed in the paper. In this algorithm, the original image is divided into different parts of blocks. Then parts of the blocks are selected to be embedded with watermark respectively according to the texture complexity. Optimal threshold is to be gained by using the iterative threshold method, and the DWT coefficients are divided into two categories by using the optimal threshold. According to the result of the classification, different methods are used to overlay the watermark signal respectively. For the main purpose of keeping balance between invisibility and robustness, the watermark embedding strength for each block is decided by fruit fly optimization algorithm. And in this way, the goal of embedding the watermark adaptively has been attained. Quantity of attack experiments is simulated, and the data show that the proposed algorithm is an effective way.Fourth, according to the present status that the geometric attacks are hard to resist, a kind of wavelet domain digital watermarking algorithm based on SIFT feature points and singular value decomposition which combines image features with transform domain method is proposed in the paper. Image stable feature points gained by the SIFT transform are used for identifying the watermark embedding positions. Then singular value decomposition and wavelet decomposition are used to embed the watermark into the carrier image. At the same time, if the feature points are used to locate and extract the watermark, the goal of effectively resisting geometric attacks can be achieved. It can be concluded that the algorithm is viable in some degree by analyzing the extracted watermark and experimental data. |
PDF Full Text Request