摘要
研究缩小图像的防伪水印问题.针对传统三维网格防伪水印算法只对正常拍摄的图像效果好,对缩小图像加入水印后,防伪水印的不可见性和图像的抗攻击性均变弱的弊端.提出了一种考虑图像缩小后特征萎缩的防伪水印算法.对于缩小图像,由于其图像特征萎缩,需要将三维模型变换为二维数值矩阵,并采用离散余弦变换进行二维分块;然后采用粒子群优化算法找到防伪水印的最优嵌入位置,将防伪水印信息嵌入到最佳位置,最后通过逆变换得到含有防伪水印的三维网络图像.仿真结果表明,该方法嵌入的防伪水印不仅可抵抗多种不同类型攻击,而且具有良好的不可见性,比传统算法防伪准确性提高了64%,抗攻击性提高了25%,为图像的版权保护提供了有力支持,在数字图像作品版权保护中具有极强的实际应用价值.
Research the narrow image of watermark security problems.In view of the traditional three-dimensional grid anti-counterfeit watermarking algorithm for normal images taken effect is good,to reduce the image watermarking,anti-counterfeiting watermark invisibility and image of the disadvantages of aggression are weaker.Is suggested to consider an after shrinking anti-counterfeiting watermark algorithm of reduction characteristics.To reduce images,shrinking due to its image characteristics,needs to be 3dmodel transformation for 2-d numerical matrix,and USES the two-dimensional block discrete cosine transform;And then USES the particle swarm optimization algorithm find the optimal anti-counterfeiting watermark embedding positions,to embed the anticounterfeiting watermark best position,finally with anti-counterfeiting watermark is obtained by inverse transformation of three-dimensional network image.Simulation results show that the method of embedded anticounterfeiting watermark not only can resist a variety of different types of attacks,and has good invisibility,anticounterfeiting accuracy increased by 64%than the traditional algorithm,resistance to attack increased by 25%,for image copyright protection provides a strong support,in the work of digital image copyright protection with strong practical application value.
出处
《微电子学与计算机》
CSCD
北大核心
2014年第10期151-155,共5页
Microelectronics & Computer
基金
四川省教厅青年基金项目(12ZB053
13ZB0105)
关键词
缩小图像模型
防伪水印
离散余弦变换
几何攻击
Reduce the image model
Anti-counterfeiting watermark
Discrete cosine transform
Geometric attacks
