Research On Algorithm For Improving The Capacity Of Digital Watermarking

Digital watermrking technique, a new method for copyright protection,has become one of the hotspot techniques in the multimedia security field in recent years, and the basic idear is to embed secret information into the digital data in order to protect the property rights of the owner.At present,the algorithm in the transform domain is one of the hotspot techniques,such as frequency and wavelet domain. The theory of wavelet transform is a new branch of mathematics appeared in recent years. Because the wavelet transform has good local properties, the wavelet based digital image watermarking technology much attention and become a research hotspot.There are many aspects to be noticed in watermarking design, for example, imperceptibility, robustness, security and capacity. Image watermarking capacity is an evaluation how much information can be hidden in a digital image. Embedding ca-pacity and the invisibility of the watermark is a contradiction. How to design a new embedding algorithm, which can both improve invisibility and watermark capacity, has always been a challenging issue. In this paper, the algorithm improving the water-mark capacity is studied. The main content and contributions are:1. Study the basic principles of digital watermarking technique, the basic frame-work and typical algorithms and the corresponding performance evaluation criteria. I expatiate on the main theory of wavelet transform and typical algorithms of digital image watermarking;2. Analysis of the relationship between the invisibility, robustness, and the emb-edding capacity in digital watermarking. Analysis of the channel capacity of waterm-arking system from the perspective of information theory. Maximum watermark based on the content of image is cited into wavelet domains to calculate the capacity of a digital image watermarking. Analysis and calculation of the embedding capaci-ty on the spatial and wavelet domain;3. Simulation analysis of the relationship between the wavelet coefficients and the image texture. With human visual model, proposed an adaptive watermarking algorithm based on tolerance to noise of wavelet coefficients in wavelet domain. Simulation results show that the algorithm can improve the watermark embedding capacity, and this algorithm has certain advantages in its robustness and security.