基于弱纹理块的扩频隐写分析

Analysis of spread spectrum steganography based on weak textured patch

提出了一种基于弱纹理块的扩频隐写分析方法,可以对图像扩频隐写强度进行有效估计。该方法根据扩频隐写过程中的图像灰度截断效应对图像中各图像块进行预筛选,避免灰度截断干扰。然后基于图像块的梯度统计特性从扩频隐写图像中选取弱纹理块,并利用所选弱纹理块根据主分量分析方法估计图像中秘密信息的嵌入强度。仿真实验表明,该方法能很好的估计扩频隐写图像中秘密信息的嵌入强度。

To estimate the strength of this paper proposes a spread-spectrum st sequential spread-spectrum steganography in images effectively, eganalysis scheme based on weak textured patch. Considering the grayscale threshold effect occurred in the process of steganography in grayscale images, it proposes a scheme to mark the patches which may have been rounded. So when it selects weak textured patches from a single noisy image based on the gradients of the patches, it may avoid selecting the patches. Then it estimates the strength of the steganography based on weak textured patches using principal component analysis （PCA）. The experimental results show that the proposed spread-spectrum steganalysis scheme achieves better estimation than the state-of-the-art algorithms.