中文摘要：

针对压缩域音频隐写算法嵌入容量较小、算法较为复杂、不可感知性较差的问题,提出一种基于压缩域的内嵌型自适应隐写算法。根据信息嵌入的多少自适应选取嵌入位置,依据秘密信息的奇偶性与码表序号的奇偶性,按照规则修改码表序号实现信息嵌入,提取信息时仅需要提取帧边信息找到使用的码表序号。该算法在容量方面较经典的MP3stego算法提升约300%,算法复杂度降低,信息的嵌入提取速率增加,对不可感知性影响较小。实验结果表明了该算法的可行性、实时性与实用性,其能够满足安全通讯的需要。

英文摘要：

Aiming at the problems of smaller embedding capacity,higher complexity and poorer imperceptibility of compressed domain audio steganography,a kind of self-adaptive inlaid-type algorithm based on compressed domain was proposed.Embedding position was selected according to the secret information,and the rule was used to modify the code table numbers to realize information embedding based on the parities of the secret information and table number.When extracting secret information,only the frame edge information was needed to find the numbers of used code tables.On the aspect of capacity,the algorithm lifts about three times than classical MP3 stego algorithm,the complexity is reduced,embedding rate is increased and it has smaller influence on imperceptibility.The experimental results verify the feasibility,timeliness and practicality of the algorithm,which satisfies needs of secure communication.