摘要
针对利用MP3边信息中主数据起始位置值作为特征的隐写分析方法存在嵌入少量秘密信息时检测率低的问题进行了改进,将MP3Stego隐写前后的数据帧主数据起始位置差扩大,以减小特征提取时重压缩估计方法带来的误差,从而提高了音频隐写分析检测率.实验结果表明,该算法对MP3Stego音频隐写可进行有效的隐写分析与检测,尤其是在嵌入少量秘密信息时检测率较现有算法有明显提高.
An improved method for MP3Stego was present to solve the problem of the low accuracy when the secret message was lit- tle, and the algorithm was based on the feature of the main-data-begin in the side information. The difference of the main-data-begin between the cover and the steganography audio embedded by MP3Stego had been enlarged. The error which was brought by re-com- pression estimate bad been decreased when the features were gained, and the accuracy of audio steganalysis was improved. Experimen- tal results indicate that the proposed algorithm can detect the MP3Stego effectively and can achieve better detection performance than the existing algorithm when the secret message is little especially.
出处
《小型微型计算机系统》
CSCD
北大核心
2015年第3期572-575,共4页
Journal of Chinese Computer Systems
基金
国家自然科学青年基金项目(61309022)资助
陕西省自然科学青年基金项目(2013JQ8031)资助