基于奇偶量化的彩色图像盲水印算法

BLIND WATERMARK FOR COLOUR IMAGES BASED ON PARITY QUANTIZATION

提出一种基于奇偶量化的彩色图像小波域的数字水印算法。首先对水印图像进行置乱处理,然后对图像进行翻转,最后利用混沌系统产生的混沌序列对水印进行加密。水印嵌入时,采用小波变换并利用奇偶量化规则将加密后的水印嵌入到载体图像蓝色分量最深层的四个子带中。水印提取时,不需要原始图像的参与,实现了盲提取。仿真实验和性能分析表明,该算法有很好的不可见性和鲁棒性,可以有效地抵抗JPEG压缩、噪声、剪切等攻击。

The thesis proposes a parity-quantization-based colour image wavelet domain digital watermarking algorithm.The algorithm scrambles the watermark image at first,then flips it over,finally encrypts the watermark with chaotic sequences generated by the chaotic system.At embedding,wavelet transform and parity quantization discipline is used to embed encrypted watermark into the deepest four sub-bands of the host image＇s blue branch.At extraction,blind extraction is realised that does not require the participation of the original image.Emulation experiments and performance analysis validate the invisibility and robustness of the algorithm,as well as its resistance against attacks such as JPEG compression,noise,cutting and so forth.