一种透明健壮大容量的音频密写算法  

A TRANSPARENT ROBUST AND HIGH CAPACITY AUDIO STEGANOGRAPHY ALGORITHM

作　　者：李跃强[1] 丁超[1] 孙星明[2] 

机构地区：[1]怀化医学高等专科学校,湖南怀化418000 [2]湖南大学计算机与通信学院,湖南长沙410082

出　　处：《计算机应用与软件》2010年第2期79-80,168,共3页Computer Applications and Software

基　　金：国家自然科学基金(60373062);教育部科研重点项目(03092);湖南省教育厅科研基金(05C020)

摘　　要：提出一种新的基于离散余弦变换的大容量数字音频密写算法。音频的低频DCT系数集中了音频的重要信息,而高频DCT系数只包含少量信息。根据这一特性,采取不改变、缩小及增加高频DCT系数的方法,在高频DCT系数中嵌入秘密信息。实验证明,该算法具有较大的嵌入量、较强的健壮性、较好的透明性,提取属盲提取,能经受重采样、重新量化、添加噪声、低通滤波、音频格式转换等常见信号处理及攻击。This paper proposes a new algorithm of digital audio steganography with high capacity based on discrete cosine transform. The important information of the audio is contained in the low frequency coefficient of the DCT, just a bit of information is contained in its high frequency coefficient. According to the features,of the audio, secret information can be embedded into the high frequency coefficient of the DCT while the coefficient either remains unchanged, increased or reduced. Experiment demonstrates that the algorithm has preferable high embedding capacity, is robust, fairly transparent and blind against common signal processing manipulations and attacks, such as re-sample, demodulation, noise, low-pass filter and audio format conversion, etc.

关 键 词：音频 密写 离散余弦变换 透明 健壮 盲提取 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置] TN912.3[自动化与计算机技术—控制科学与工程]