摘要
提出一种基于神经网络检测的自同步隐蔽信息传输系统,该系统满足以下特点:利用人的听觉系统分辨率限制实现隐藏信息伪装性传输,能有效克服传统加密通信易被攻击者监听、截取的缺点;先应用神经网络提取嵌入到语音中的数据,再进行相关检测确定自同步信号的存在。与直接提取数据后进行相关检测的方法相比,该方法能有效降低比特误码率,从而降低漏检和虚检概率,使隐藏信息的准确传输率明显提高。
A Transporting System of Hiding Information with serf-synchronous signal detected by neural network is presented in this paper. It has two advantages. One is it not only avoids the traceable and vulnerable disadvantages of the conventional manner, but also prevent attacker from stealing the hiding information validly. The other is it employ neural network to extract the self-synchronous signal and detects it through correlation coefficient. Compared with the manner of directly extracting and detecting it according to the correlation coefficient, the new way make the error bits ratio, the error detecting probability, and leak detecting probability was decreased validly. As a result the correct transporting ratio of hiding information is increased clearly.
出处
《微电子学与计算机》
CSCD
北大核心
2007年第3期142-144,149,共4页
Microelectronics & Computer
基金
国家自然科学基金项目(60172067)
教育部博士点基金项目(20020558038)
广州市重大科技攻关项目(2000-Z-004-01)
关键词
神经网络
自同步
数字水印
隐藏信息
相关检测
加密通信
neural network
synchronous signal
digital watermark
hiding information
correlation coefficient encryption communication
