主要对基于图像的广义隐写分析算法进行了研究,选取4种对信息隐藏分析效果最好的图像质量度量标准(Image Quality Metrics),运用误差反传(Back Promulgate)神经网络设计了一套广义隐写分析系统,并将其用于最终隐写图像的检测分类上.经实验证明,该方法具有较高的准确率和普遍的适应性,因而可用于实际的隐写图像检测分析中.
Mainly focuses on the universal steganalysis based on image. Using BP neural networks, 4 specific IQMs that are most consistent and accurate vis- 6- vis the effects of steganography are identified, and a universal steganalysis system is deviced to classify the stego images. Experiment results indicate that the method has a good detection rate and a wide range of application, so this can be applied to the real world.