一种基于矩阵奇异值分解和奇异值量化的数字水印算法

The Digital Watermarking Based on the SVD and Singular Value Quantification

在图像矩阵的奇异值中嵌入隐秘信息,这样的嵌入算法鲁棒性很强,但嵌入的信息量却很有限.本算法是将图像矩阵分解为若干小的矩阵序列,使用将每个小矩阵的奇异值进行量化的方法嵌入隐秘信息,这样既增强了算法的鲁棒性,也提高了算法的信息隐藏量.实验结果表明该算法可行,并且可以抵抗多种攻击.

An algorithm based on singular value decomposition （SVD） is proposed, which hides secret information in singular value vector. The singular value vector is the inherent character of matrix; it can improve the robust of the algorithm that hides secret information in singular value vector. The result of experiment indicates that the algorithm is feasible and robust.