基于人眼对比敏感视觉特性的彩色图像盲水印算法

为了解决水印透明性、鲁棒性与水印容量之间的平衡问题,结合人眼对比敏感度视觉特性,提出了一种彩色图像的盲水印算法。算法首先结合人眼对比敏感度视觉特性及其模型,在频域中计算人眼感知原始图像的对比敏感度阈值;其次结合人眼对强度的感知模型,计算水印图像信息的强度感知值;然后结合嵌入强度系数值,以感知原始图像的对比敏感度阈值为水印嵌入量的门限,实现数字水印的嵌入;最后依据密钥,采用逆过程提取水印。实验结果表明,在压缩质量因子为20%、中间剪切1/4区域的较强攻击下,提取水印与原水印的相似度达到0.78以上,且保证了较好的水印透明性。

Blind watermarking algorithm for colour image based on human visual contrast sensitivity characteristics

Aiming at the balance between watermarking transparency, robustness and watermarking capacity, a blind watermarking algorithm for color images is proposed based on the contrast sensitivity characteristics of human visual system (HVS) in this paper. In this algorithm, firstly, according to the contrast sensitivity characteristics of HVS and its calculation models, the contrast sensitivity thresholds from HVS perceiving the original image are calculated in the frequency field. And combined with the strength perception model of HVS, the perceived strengths of the watermarking images are calculated. Then according to the coefficients of embedding strengths, the digital watermarkings are embedded by taking the contrast sensitivity perceived original images as the thresholds of the watermarkings embedding quantity. Finally, based on the key, the inverse process is carried out to extract the watermarkings. The experimental results show that under the strong attack that the watermarked image is compressed by quality factor 20% and 1/4 size is sheared in the intermediate region of image, the similarity between the original watermarking and the extracted watermarking can reach more than 0.78. And the proposed watermarking algorithm has good transparency.