【摘要】 为了提高水印的鲁棒性和减小对应版权信息的大小,本文提出了一种基于Arnold变换、DWT和SVD的零水印算法。该算法将载体图像进行DWT变换,对低频子带进行分块SVD;比较置乱后的最大奇异值的大小,若大于则对应位为1,否则为0;将结果与预处理过的水印做异或运算,得密钥矩阵。该矩阵为检测水印提供密钥。实验证明,本算法对图像压缩、滤波、模糊、噪声及脏图具有良好的鲁棒性。更多还原

【Abstract】 In order to improve the robustness of the watermarking and reduce the size of the copyright information,this paper presents an improved non-watermarking algorithm based on DWT and SVD.When embedding,performs DWT on the host image to decompose it into four non-overlapping sub-bands,then selects the LL sub-bands to perform block SVD.Then the largest singular value of each block is selected to form a matrix.The values of adjacent element of scrambling matrix are compared,when the compared value is true,the corresponding value is 1,or is 0.The key matrix is formed when doing XOR operation on the result and the scrambled and extended watermarking image.The key matrix can be used to detect the watermarking image later.The experimental results show that this algorithm has good performance to the Compression、filtering,noise,fuzzy and dirty pictures.更多还原