Detect Hugo By Merging Feature Sets
(0) 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.
watermarking model layered information watermark property
Qingxiao GUAN Jing DONG
Department of Automation, University of Science and Technology of China National Laboratory of Patte National Laboratory of Pattern Recognition, Institute of Automation
国内会议
北京
英文
260-264
2012-04-01（万方平台首次上网日期，不代表论文的发表时间）