基于非负矩阵分解的图像脆弱水印算法

摘要

提出一种基于非负矩阵分解NMF（Non—negative Matrix Factorization）的脆弱数字水印算法。算法利用用户密钥构造NMF基矩阵，并在图像NMF分解过程中保持不变，二值水印图像嵌入NMF分解系数矩阵。实验结果本算法具有较强的鲁棒性，同时用户密钥保证的算法的脆弱性。%This paper presents a digital watermarking algorithm based on non-negative matrix factorization （NMF）. The partial representation--based coefficient matrix of a carrier image is constructed by the NMF first, and the grey watermarking image is embedded to the coefficient matrix. Then the base matrix of NMF is utilized as a key to extract the watermarking. The new NMF2 based algorithm is compared with the DCT-based algorithm in correlative experiment in order to illustrate its effectiveness. The results show that the former is better than the latter in anti-cut and anti-filter condition, and is as good in other attack conditions.