摘要
提出一种基于HMT预测的ICA盲源分离主动分析方法,该方法仅仅需要得到一幅隐秘图像,并采用小波域HMT(hidden markov tree)模型预测出载体图像的一个估计,最后使用ICA盲分离来提取秘密信息和载体图像。仿真实验给出了该方法提取出的载体图像和秘密信息,并对其进行了分析。
Two copies of stego image are needed in the active steganalysis proposed by Chandramouli, as an improvement, a method of active steganalysis is proposed in which only one copy of stego image is needed and the practicability is enhanced. An estimate of a cover image by wavelet-domain hidden Markov tree（HMT） is obtained; and the estimate is regarded as amother stego-image which is different in the embed ratio. Then cover image and the secret message are separated by using the independent component analysis（ICA）. Simulation experiments validate the advantage of HMT model, and give the results of blind separation.
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2008年第10期1059-1061,1066,共4页
Geomatics and Information Science of Wuhan University
关键词
主动隐写分析
HMT模型
ICA
仿真
active steganalysis
HMT model
independent component analysis
simulation
