基于人类视觉特性的自适应水印嵌入方案

Adaptive image watermarking technique based on human visual system

针对已有算法在嵌入量和嵌入强度自适应控制方面存在的问题,结合人类视觉对纹理区域的感知特性,通过噪声可见函数(NVF)对图像块复杂度的计算,提出了一种在DCT直流系数进行嵌入的水印算法。该算法能够根据图像块复杂度计算结果实现嵌入强度的自适应控制。实验结果表明,所提算法相比已有算法提高了嵌入容量和嵌入强度的自适应控制程度,并且一定程度上避免了由于DC系数嵌入数据所出现的块效应现象,对图像处理和常见的图像攻击也具有很好的稳健性。

In order to improve the capacity and self-adaptive control of embedding intensity,this paper proposed a new method for embedding information into DC coefficients of DCT-domain,which combined noise visibility function（NVF） and human visual system which was sensitive to texture features.The algorithm could control the embedding intensity self-adaptively,according to block complexity of images computed using NVF.Experimental results show that the proposed algorithm improves the capacity and self-adaptive control of embedding intensity,and to some extent,avoids the block effect caused by the embedding information.Additionally,it is robust against common image processes and attacks.