论文部分内容阅读
Hugo is a recently developed steganographic method. So far it is believed to be a very secure steganographic-method. In this paper we focus on detecting Hugo by merging features. We combine 4 effective feature sets to get better performance and use Support Vector Machine(SVM) as classifier. Experimental results show that combined features can detect Hugo effectively. We also take the advantage of subspace learning to further investigate how to reduce the dimension of the combined feature set and finally conclude our effective components of detecting for breaking Hugo.