基于特征子空间的数字水印技术

根据特五子空间的原理，提出了一种新颖的基于特征子空间的数字水印算法，水印可以嵌入到不同的子空间中从而可以发展为多水印算法，由于特征子空间之间的正交特性，不同子空间中的水印相互独立，其稳健性不随水印数目的增加而变化，水印强度参数根据视觉系统的照度掩蔽和纹理掩蔽特性适当地进行调整，从而同时保证二值水印图像的稳健性和不可见性，该文就单水印，双水印以及多水印情形进行了仿真并进行JPEG压缩，图像缩放，剪切等失真测试，实验结果表明该算法具有很好的感知质量和稳健性并验证了以上分析。

A Digital Watermarking Algorithm Based on Subspace

As an effective technique of copyright protection, digital watermarking techniques have become a new focus of interest in recent years. In this paper, a novel digital watermarking algorithm is proposed based on the subspace principle. The watermark can be simultaneously embedded into several subspaces thus to become a multi watermark scheme. Due to the orthogonality between different subspaces, the watermarks in different subspaces are independent whose robustness remains unchanged even as the number of subspaces used for watermarking increases. According to the visual mask effect of human vision system, the watermark power control factors are properly selected to make watermark power under the just noticeable threshold, the relation between robustness and imperceptibility can be well balanced. We test the algorithm performance in the cases of single watermark, double watermarks and multi watermarks when different distortions such as JPEG compression, image resizing and cropping are applied. Experimental results show good results on image perceptual quality and watermark robustness, which verify the correctness of our analysis.