中文摘要：

在研究Costa设置和Cox高斯分解的基础上，根据信源的随机状态，提出了一种新的信息隐藏传输技术。通过信源Laplacian条件分割，并根据图像系数方差的不同分布区域对高斯信道进行并行分解，该技术实现了在二义性攻击条件下的一种稳健信息隐藏传输方法。结合应用SS和QIM（DC—DM）技术，实验结果表明，该技术在信息嵌入量、传输码率和信道容量之间都具有很好的折中性能。

英文摘要：

Based on Costa set-up and Cox Gaussian-decomposition for stochastic mixture source, a novel information-hiding communication technology is proposed . By means of splitting to the source with Laplacian distribution , the method forms a Markov chain between the global and local statistics of image coefficients. On the other hand, though parallel Gaussian channel according to its variances intervals, the system approaches a good transmission performances under various WNR that corresponds to the attack channel ambiguity. Comparing the theoretical resuits with the empirical data obtained by experimentation considered SS and QIM （DC-DM） techniques, the system resolves the trade-off between embedding capacity and channel rates as well as robust.