引用本文
  • 蔡森,任延珍,王丽娜.基于联合失真的AAC安全隐写算法[J].信息安全学报,2022,7(2):1-15    [点击复制]
  • CAI Sen,REN Yanzhen,WANG Lina.An AAC Steganographic Algorithm Based on Joint Distortion[J].Journal of Cyber Security,2022,7(2):1-15   [点击复制]
【打印本页】 【在线阅读全文】【下载PDF全文】 查看/发表评论下载PDF阅读器关闭

←前一篇|后一篇→

过刊浏览    高级检索

本文已被:浏览 1396次   下载 611 本文二维码信息
码上扫一扫!
基于联合失真的AAC安全隐写算法
蔡森, 任延珍, 王丽娜
0
(空天信息安全与可信计算教育部重点实验室, 武汉大学国家网络安全学院 武汉 中国 430072)
摘要:
随着互联网技术的普及, 越来越多的音视频通信应用融入到了人们的日常生活中。 AAC(Advanced Audio Coding), 作为目前互联网应用中使用最广泛的音频压缩编码标准之一, 拥有优秀的压缩效果和出色的音频质量, 使得越来越多的音视频作品利用 AAC 进行编码传输, 这也为信息隐藏提供了新的、更多、更好的隐写空间。本文分别对现有的隐写算法生成的含密音频的Huffman 码字进行统计分析和算法中基于掩蔽曲线理论设计的失真函数进行分析, 发现现有隐写算法会在不同程度上破环音频的 Huffman 码字的统计安全性和音频的听觉掩蔽性。为了解决现有的隐写算法在统计安全性和听觉隐蔽性方面的不足, 本文提出了基于 Huffman 码字统计分布特性和实际听觉掩蔽阈值曲线相结合的 AAC 隐写联合失真代价函数。同时, 利用最小失真编码框架 STCs(Syndrome-Trellis Codes)实现了面向 AAC 压缩参数域的自适应安全隐写算法。 实验结果表明, 本文提出的隐写方法的隐写容量在理论上最大可达 12 kbps。另外, 在统计安全性方面, 相较于现有的隐写方法, 最高可以提升 30%。在听觉质量方面也有进一步的提升。 在可扩展性方面, 由于 MP3(Moving Picture Experts Group Audio Layer III)的编码流程和 AAC 的编码流程在大体上类似, 本文所提出的新方案可以经过微小的调整后, 便可直接应用在基于 MP3 的隐写算法中, 对现有基于 MP3 的隐写算法在安全性和音频质量方面也有一定的提高。
关键词:  AAC  Huffman 编码  隐写术  听觉掩蔽效应
DOI:10.19363/J.cnki.cn10-1380/tn.2022.03.01
投稿时间:2020-12-10修订日期:2021-03-02
基金项目:本课题得到国家自然科学基金(No.61872275,No.U1836112,No.61876134)资助。
An AAC Steganographic Algorithm Based on Joint Distortion
CAI Sen, REN Yanzhen, WANG Lina
(Key Laboratory of Aerospace Information Security and Trusted Computing, Ministry of Education, School of Cyber Science and Engineering, Wuhan University, Wuhan 430072, China)
Abstract:
With the popularization of Internet technology, more and more audio and video communication applications have been integrated into People’s Daily life. AAC (Advanced Audio Coding), as one of the most widely used Audio compression Coding standards in Internet applications, has excellent compression effect and excellent Audio quality, making more and more Audio and video works use AAC for Coding transmission. This also provides new, more and better steganographic space for information hiding. In this paper, the statistical analysis of Huffman code words with dense audio generated by the existing steganography algorithm and the distortion function designed by the algorithm based on the masking curve theory are carried out respectively, and it is found that the existing steganography algorithm will destroy the statistical security of Huffman code words and audio masking to varying degrees. In order to solve the shortcomings of existing steganography algorithms in statistical security and auditory concealment, the AAC steganographic joint distortion cost function is proposed based on Huffman code word statistical distribution characteristics and actual auditory masking threshold curve. At the same time, an adaptive steganographic algorithm for AAC compression parameter domain is implemented by using the minimum distortion coding framework STCs (Syndrome-Trellis Codes). The experimental results show that the maximum steganographic capacity of the proposed steganographic method can reach 12 kbps. In addition, statistical security can be improved by up to 30% compared to existing steganographic methods. There was also a further improvement in auditory quality. In terms of scalability, because the encoding process of MP3 (Moving Picture Experts Group Audio Layer III) is basically similar to that of AAC, the new scheme proposed in this paper can be directly applied to steganographic algorithm based on MP3 after minor adjustment. It also improves the security and audio quality of the steganographic algorithm based on MP3.
Key words:  AAC  Huffman Code  steganography  masking effect of the psychoacoustic model