一种针对JPEG图像的通用隐写分析算法

摘要

提出了一种针对JPEG图像的通用隐写分析算法。该算法提取了15个具有良好分类特性的特征参数，输入构建的KS—SVM分类器．以达到检测载密图像的目的。实验结果表明，该算法的检测正确率较高，检测速度快，能够实现针对各类JPEG载密图像的有效检测。%A blind steganalysis algorithm for JPEG images was proposed. The algorithm calculates a total of 15 characteristic parameters from each JPEG image, and using LS-SVM classifier to classify cover and stego images, which can apply to blind steganalysis. The experiments results show that, the algorithm has good performance in both detection rate and computation speed, and it is also reliable to steganalysis for all kinds of JPEG images steganography.