量化水印算法分析

Analysis of quantization-based watermarking

对量化水印算法稳健性与抗攻击能力问题,给出了公有水印算法稳健性充分必要条件和抗攻击能量上限,弥补了抗攻击能力评估上没有精确界限的缺陷。分析了载体系数的改变幅度及固定保真度优化中附加模板的取值范围,部分解决了保真度优化中出现的保护水印问题。指出了量化信息编码与一般变量信息承载之间的区别,明确了在信息容量计算中的差别原因。构造了优化视觉效果的嵌入方法和与其等价的其它10个嵌入方法。在同等的载体改变范围内,提出了比基于mod运算的水印嵌入方法抗攻击能力高一倍的量化水印算法,弥补了基于mod运算的水印嵌入方法的抗攻击能力比较低的缺陷。构造了嵌入灰度值的水印算法,避免了量化嵌入算法中把灰度值转换为二进制嵌入后所带来的缺陷。实验验证了此算法及理论分析结果。

Necessary and sufficient conditions for robustness to the public watermarking algorithm and its upper limit of attack-resistance energy were given, which supplied a gap that there was no exact upper limit in the attack-resistance evaluation. Then it analysed change amplitude of the carder＇s coefficients and range of the added pattern in fidelity-fixed optimization, which partly solved the problem of protecting watermarking during fidelity-fixed optimization, and pointed out the difference between quantization encoding and common variable information payload and the reasons of difference in information capacity calculation. After that, a vision optimized embedding algorithm and other ten equivalent ones were constructed. And then a quantization watermarking scheme with twice attack-resistance of mod-based embedding algorithms was proposed in the same carrier altered range, which overcomed the disadvantage of low attack-resistance to Mod-based watermarking schemes. Also, it constructed gray level embedding schemes which avoided the transformation from gray value to binary. Finally, the experimental results support the theoretic analysis and presented schemes.