Abstract: Aiming at the limitation of watermark capacity that traditional Singular Value Decomposition(SVD) watermark algorithm shows, this paper proposes an image watermark algorithm based on SVD. The original image is divided into blocks, and several former Singular Values(SVs) are extracted from each block as the image characteristics, which are used to construct zero-watermark. Simulation experimental results show that the algorithm can reduce caculation complexty, increase the watermark capacity, and perform robustly.

Key words: zero-watermark, Singular Value Decomposition(SVD), image characteristic, watermark capacity

摘要： 针对传统的基于奇异值分解的水印算法水印容量较低的缺陷，提出基于奇异值分解的零水印算法，在分块的基础上提取奇异值向量的前若干项，将其作为图像特征，并利用此特征进行零水印构造。仿真实验结果表明，该算法能降低计算复杂度，增加水印容量，稳健性较好。

关键词: 零水印, 奇异值分解, 图像特征, 水印容量

CLC Number: 

• TP309.2