数字水印信道模型的快速估算

Fast Estimation of Watermarking Channel Modeling

近年来,数字水印技术开始广泛用于数字图象的版权保护并提出了大量的数字水印方案。在本文中分析了一种用于静态图象版权保护的类似于扩谱技术的DCT变换域的数字水印信道模型,在JPEG算法中, DCT变换是作用在8×8大小的像素分块上的。数字水印通过编码信息来跟踪电子产品的非法使用。我们用广义高斯分布去作为自然界图像经过DCT变换后的分布模型, 作为我们工作的结果, 信道模型的快速估算表达式被推导并和实验结果进行了比较。

In recent years,digital watermarking has been widely used in copyright protection for digital images,and a large number of schemes have been proposed. In this paper,a spread-spectrum-like discrete cosine transform domain (DCT domain) watermarking channel modeling for copyright protection of still digital images is analyzed. The DCT is applied in blocks of 8X8 pixels as in the JPEG algorithm. The watermark can be used to encoded information to detect illegal misuses. According to this model, Generalized Gaussian distribution is used to statistically model the DCT coefficients of different natural images. As a result of this work, fast estimating expression for channel modeling are derived and then verified by experimental results.