小波包子带频域矩和遗传算法的图像隐写分析

针对现有典型的通用隐写分析方法提取特征时多数对图像进行正交小波分解中的不足,提出了一种新的利用小波包分解和遗传算法的高效的隐写检测方案.首先对原灰度图像及其预测误差图像运用小波包分解生成多个子带,提取各子带系数的直方图特征函数多阶矩作为特征,将支持向量机的分类效果作为适应度函数值返回,指导遗传算法搜索最优的特征选择方案.针对Outguess、F5、MB1隐密图像与载体图所组成的混合图像库上的仿真实验结果表明:该方法比现有的隐写分析方法,平均检测率提高了约1%-17.8%.

Steganalysis based on wavelet packet transform and genetic algorithm

Against the shortness of general steganalysis method which use the orthogonal wavelet to extract the characteristics, a effective steganalysis method based on wavelet package decomposition (WPD) and genetic algorithm was designed. Firstly, by three scales of WPD, image and its prediction-error version were decomposed into some coefficients subbands, the multi-order characteristic function (CF) moments of histogram were extracted from these subbands. Then genetic algorithm (GA) was employed to search and identify the potential informative features combinations for classification. The classification accuracy from the support vector machine classifier is used to determine the fitness in genetic algorithm. The experiments works on mixed image library, constituting by Outguess, F5, MB1 stego-images and cover images. Results show the method can detect stego and original images reliably, and the average detection accuracy of this method exceeds those of its closest competitors by at least 1.7% and up to 17.8%.