Research On Adaptive Image Watermarking

With the rapid development of Internet and multimedia technology, digital watermarking is developed for the applications of digital multimedia, such as copyright protection, image authentication, and proof of ownership.In this paper, we firstly introduce the background of digital watermarking, main algorithms, and properties. Then image watermarking techniques and wavelet transform are described. Based on these analysis, a new watermarking algorithm based on DWT is presented. This kind of digital watermarking is blind, robust and adaptive. The wavelet coefficients of the host image are divided into blocks, in each block, the mean and difference between adjacent coefficients are calculated. The quantization steps depend on the differences. In detail, the mean values and differences are used to divide the coefficients into important groups and less important ones. Since the coefficients of smooth areas in an image are less than that of complexity areas, thus the coefficients of complexity areas can be modified largely without significant notice by eyes. That is to say, only the important groups are selected to embed watermark and their mean values are quantized. Furthermore, each coefficients in the block are modified according to the mean and their comparative magnitudes. Simulation result shows that the embedded watermark can be adaptive to the host image , and the requirements of transparency and robustness can both be achieved with good noise performance.