一种基于混沌和独立成分分析的自适应多重数字水印算法

采用混沌和独立成分分析技术，提出了一种在小波域自适应嵌入多重数字水印的算法。首先根据一个混沌标签序列从原图像中抽取一子图像，并对其进行小波分解，然后将多个互相独立的混沌水印信号依次嵌入到子图像的中频小波系数中，并自适应确定嵌入强度。根据Newman-Pearson准则检测水印是否存在，若存在，采用独立成分分析的方法提取出水印并进一步判断是否为原始水印。水印的检测与提取过程都不需要原始图像。仿真结果表明该算法对各种常见的攻击有着良好的鲁棒性，有效地解决了数字产品的多著作权问题。

Multiple adaptive digital watermarking method based on chaos and independent component analysis

Based on chaotic sequences and independent component analysis(ICA),a novel multilevel image watermarking method was proposed in wavelet domain.Firstly,discrete wavelet transform(DWT) was applied to one subimage extracted from the original image according to one chaotic sequence label and multiple independent chaotic watermark signals were generated and embedded into the middle wavelet frequency coefficients adaptively.The watermarking was detected according to the Newman-Pearson criterion without the original image.Then ICA was applied to extract the watermarking if the watermarking are the original watermarking.Situational results demonstrate that this algorithm is an effective digital watermarking method with very outstanding robustness against many common attacks such as noise attacks,geometric operations and JPEG compression,and it solves the problems of multi-copyright claim in digital products.