音频与图像相比具有信息冗余大、随机性强的特点,在音频中实现无误码的信息提取的难度更大。提出一种基于DCT域QIM(quantization index modulation)的音频信息伪装算法,算法特点如下:应用QIM原理,以量化的方式嵌入信息,根据量化区间与...音频与图像相比具有信息冗余大、随机性强的特点,在音频中实现无误码的信息提取的难度更大。提出一种基于DCT域QIM(quantization index modulation)的音频信息伪装算法,算法特点如下:应用QIM原理,以量化的方式嵌入信息,根据量化区间与信息比特的映射关系提取信息,可实现盲提取;采用改进的QIM方案,针对信息提取的误码,在嵌入端与提取端进行容错处理,保证了隐藏信息的强顽健性;隐藏容量大,可达357.6bit/s。实验表明,算法与传统QIM方法相比具有更好的不可感知性,100%嵌入的载密音频的信噪比在30dB以上,并且对于MP3压缩、重量化、重采样、低通滤波等攻击具有很强的顽健性,同时算法运算量小,易于实现,实用性强。展开更多
The weakness of Human Auditory System (HAS) led the audio steganography process to be used in hiding data in the digital sound. Audio steganography is implemented here by using Least Significant Bit (LSB) algorithm to...The weakness of Human Auditory System (HAS) led the audio steganography process to be used in hiding data in the digital sound. Audio steganography is implemented here by using Least Significant Bit (LSB) algorithm to hide message into multiple audio files. This is achieved by 1st, 2nd, 3rd, and 4th bits hiding ratios. In comparison to other used bits, hiding results show that the use of 1st bit in LSB method for embedding data is much better than those used bits as expected. In addition to that and according to the results, file’s size affects strongly upon the effectiveness of the embedding process while hiding starting position doesn’t affect upon the variation of the adopted statistical estimators regardless to which bit is used. Among the statistical estimators that have been adopted here, the Mean Absolute Error (MAE) seems to be the best one in testing hiding process.展开更多
现有音频隐写软件大多基于最低比特位(LSB)隐写,对LSB隐写的检测算法研究具有重要意义。借鉴图像隐写检测中相关位平面和相邻像素相关性的分析思想,结合16 bit wav音频LSB的隐写特性,通过对音频数据相邻向量对的奇偶和大小分情况讨论,...现有音频隐写软件大多基于最低比特位(LSB)隐写,对LSB隐写的检测算法研究具有重要意义。借鉴图像隐写检测中相关位平面和相邻像素相关性的分析思想,结合16 bit wav音频LSB的隐写特性,通过对音频数据相邻向量对的奇偶和大小分情况讨论,得出隐写会使相邻向量对的统计值增大的结论。通过经验阈值的设定,实现对音频LSB隐写的有效统计检测。理论和实验证明了该算法的正确性。展开更多
文摘音频与图像相比具有信息冗余大、随机性强的特点,在音频中实现无误码的信息提取的难度更大。提出一种基于DCT域QIM(quantization index modulation)的音频信息伪装算法,算法特点如下:应用QIM原理,以量化的方式嵌入信息,根据量化区间与信息比特的映射关系提取信息,可实现盲提取;采用改进的QIM方案,针对信息提取的误码,在嵌入端与提取端进行容错处理,保证了隐藏信息的强顽健性;隐藏容量大,可达357.6bit/s。实验表明,算法与传统QIM方法相比具有更好的不可感知性,100%嵌入的载密音频的信噪比在30dB以上,并且对于MP3压缩、重量化、重采样、低通滤波等攻击具有很强的顽健性,同时算法运算量小,易于实现,实用性强。
文摘The weakness of Human Auditory System (HAS) led the audio steganography process to be used in hiding data in the digital sound. Audio steganography is implemented here by using Least Significant Bit (LSB) algorithm to hide message into multiple audio files. This is achieved by 1st, 2nd, 3rd, and 4th bits hiding ratios. In comparison to other used bits, hiding results show that the use of 1st bit in LSB method for embedding data is much better than those used bits as expected. In addition to that and according to the results, file’s size affects strongly upon the effectiveness of the embedding process while hiding starting position doesn’t affect upon the variation of the adopted statistical estimators regardless to which bit is used. Among the statistical estimators that have been adopted here, the Mean Absolute Error (MAE) seems to be the best one in testing hiding process.
文摘现有音频隐写软件大多基于最低比特位(LSB)隐写,对LSB隐写的检测算法研究具有重要意义。借鉴图像隐写检测中相关位平面和相邻像素相关性的分析思想,结合16 bit wav音频LSB的隐写特性,通过对音频数据相邻向量对的奇偶和大小分情况讨论,得出隐写会使相邻向量对的统计值增大的结论。通过经验阈值的设定,实现对音频LSB隐写的有效统计检测。理论和实验证明了该算法的正确性。
