Research On Self-adaptive Digital Image Watermarking Algorithm Based On Blocking Techniques

With the rapid development of computers and network, people's many creation and productions are stored and released in the digital form. However, digital works can be illegally copied, forged or tampered easily. In this case, copyright protection issues become very important. Digital watermarking as a very effective method for copyright protection, has become the hot spot in academic circles.On the basis of summarizing the development background and basic theoretical framework of digital watermarking technology, this paper has researched the copyright protection issues of still images, and represents two kinds of image watermarking algorithm.According to the characteristic of HVS (human visual system, HVS), both two algorithms select color images as the carrier, and use the blocking technology based on self-organizing mapping to embed watermark adapatively. The algorithm based on wavelet transform and singular value decomposition firstly transforms the original image from RGB color space to YIQ color space and extracts the component Y to performing 8×8 blocking and blocks sorting. Then for each block performs wavelet decomposition and singular value decomposition, combining with Arnold transform, embeds the 2-value watermarking sequence after scrambling to the biggest singular value of every block. The embedding intensity varies as different sorts. Experiment results show that the algorithm has the nicer transparency and robustness.The self-adaptive image watermarking algorithm based on DCT coefficients quantizing uses DCT blocking and block sorting technology. First to component Y of original image performs blocking by Self-Organizing Mapping method and discrete cosine transform, then embeds the 2-value watermark image after Arnold transform into the carrier through quantizing and modifying its DCT coefficients, and the quantizing factor varies as different sorts. The watermark can be extracted without resorting the original image. Experiment results show that high robustness can be obtained under the attack of gauss noise, salt & pepper noise, filtering, cropping and JPEG compression etc.