基于自适应奇异值调制的抗转码视频隐写算法

管萌萌; 曹纭; 张怡暄; 赵险峰

引用本文：

管萌萌,曹纭,张怡暄,赵险峰.基于自适应奇异值调制的抗转码视频隐写算法[J].信息安全学报,2018,3(6):42-54 [点击复制]
GUAN Mengmeng,CAO Yun,ZHANG Yixuan,ZHAO Xianfeng.A Transcoding-resistant Video Steganographic Algorithm based on Adaptive Singular Value Modification[J].Journal of Cyber Security,2018,3(6):42-54 [点击复制]

本文已被：浏览 5772次下载 5827次	码上扫一扫！
基于自适应奇异值调制的抗转码视频隐写算法
管萌萌^1,2, 曹纭^1,2, 张怡暄^1,2, 赵险峰^1,2
0 字体:加大+\|默认\|缩小-
(1.中国科学院信息工程研究所信息安全国家重点实验室北京中国 100093;2.中国科学院大学网络空间安全学院北京中国 100093)

摘要:

近年来，随着网络和多媒体处理技术的飞速发展，互联网视频用户数量呈爆炸性增长之势，目前视频流量已占网络总流量的一半以上，这为以视频为载体的隐蔽通信提供了很好的应用背景。但是另一方面，几乎所有的视频网站都会对上传视频进行转码处理，转码后载体视频的码率、图像组GOP （Group of Pictures）结构、量化参数QP （Quantization Parameters）等关键压缩参数都会受到无法预知的影响，这极大地限制了当前绝大多数基于压缩域嵌入的视频隐写算法的应用。要利用视频网站对应的有损转码信道进行隐蔽通信，关键在于设计能抵抗转码攻击的高鲁棒性，并具有足够安全性的隐写方法。针对这一问题，本文提出了一种面向小波变换域，结合奇异值分解SVD （Singular Value Decomposition）和量化索引调制QIM （Quantization Index Modulation）的自适应鲁棒隐写方法。该方法基于目标网站的转码特性仿真实现本地编码器，在编码的过程中完成嵌入。嵌入在视频帧亮度分量的DWT （Discrete Wavelet Transform）域进行，通过采用自适应量化步长QIM的方法对分块最大奇异值进行调制达到嵌入消息的目的。另外，提出一种间隔修正的嵌入策略，很好地抑制了嵌入引起的视频闪烁效应。实验结果表明，提出方法提升了含密载体视频经过不同程度的转码后的秘密消息生存率，同时很好地保持了视觉质量和压缩效率。

关键词: 信息隐藏视频隐写鲁棒自适应嵌入

DOI：10.19363/J.cnki.cn10-1380/tn.2018.11.04

投稿时间：2018-08-15修订日期：2018-10-11

基金项目:本课题得到中国科学院信息工程研究所基础前沿项目（No.Y7Z0371102），国家自然科学基金（No.61802393，No.61872356），国家重点研发计划（No.2016YFB0801003，No.2016QY15Z2500，No.2017YFC0822704）资助。

A Transcoding-resistant Video Steganographic Algorithm based on Adaptive Singular Value Modification

GUAN Mengmeng^1,2, CAO Yun^1,2, ZHANG Yixuan^1,2, ZHAO Xianfeng^1,2

(1.State Key Laboratory of Information Security, Institute of Information Engineering, Chinese Academy of Sciences, Beijing 100093, China;2.School of Cyber Security, University of Chinese Academy of Sciences, Beijing 100093, China)

Abstract:

Recent years have witnessed the exploding number of Internet video users as the rapid development of network and multimedia processing technologies. At present, video traffic has accounted for more than half of the total network traffic, which provides a good cover for video-based covert communication. On the other hand, almost all video websites will transcode the uploaded video. The key compression parameters such as the bit rate of the carrier video, the GOP (Group of Pictures) structure, and the quantization parameter QP (Quantization Parameters) will be transcoded. Unpredictably, this greatly limits the current majority of video steganography applications based on compressed domain embedding. To use the lossy transcoding channel corresponding to the video website for covert communication, the key is to design a steganographic method that is robust against transcoding attacks and has sufficient security.For this problem,this paper proposes an adaptive robust steganography method for wavelet transform domain combined with SVD (Singular Value Decomposition) and QIM (Quantization Index Modulation). The process is embedded in the DWT (Discrete Wavelet Transform) field of the video frame, and the maximum singular value of the block is modulated by the method of adaptive quantization step size QIM to achieve the purpose of embedding the message. In addition, an interval-embedded embedding strategy is proposed to suppress the video flicker effect caused by embedding. The experimental results show that the proposed method improves the confidential message survival rate of the video with the fixed carrier after different degrees of transcoding, and better maintains the visual quality and compression efficiency.

Key words: information hiding video steganography robust adaptive embedding