利用跨模态信息检索的鲁棒隐蔽通信  

RoCC:robust covert communication based on cross-modalinformation retrieval

作　　者：张晏铭 陈可江 丁锦扬 张卫明[1] 俞能海[1] Zhang Yanming;Chen Kejiang;Ding Jinyang;Zhang Weiming;Yu Nenghai(School of Cyber Science and Technology,University of Science and Technology of China,Hefei 230000,China)

机构地区：[1]中国科学技术大学网络空间安全学院,合肥230000

出　　处：《中国图象图形学报》2024年第2期369-381,共13页Journal of Image and Graphics

基　　金：国家自然科学基金项目(62102386,U2336206,62002334,62072421)。

摘　　要：目的 隐蔽通信是信息安全领域的一个重要研究方向,现有基于多媒体数据流构建隐蔽信道的方法,未考虑网络传输时波动产生的数据包丢失问题。本文提出一种基于跨数据模态信息检索技术的对网络异常具有鲁棒性的隐蔽通信方法,同时可以满足高隐蔽性和高安全性的要求。方法 提出了一个名为RoCC(robust covert communication)的通用隐蔽通信框架,它基于跨模态信息检索和可证明安全的隐写技术。所提方法将直接通信和间接通信两种形式相结合。直接通信通过VoIP(voice over internet protocol)网络通话服务进行,传递实时生成的音频流数据,接收方可以通过语音识别将其还原为文本;而间接通信则借助公共网络数据库进行载密数据的传输,接收方通过文本语义相似度匹配的方式来还原完整语义的载密文本数据,这有助于解决网络数据包丢失和语音识别误差导致的文本语义丢失的问题。结果 经实验测试,本文方法在协议上具有更好的通用性,相对Saenger方法在丢包率抵抗能力方面提高了5%,所用隐写算法满足可证安全性。同时,RoCC的数据传输率有73~136 bps(bit per second),能够满足实时通信需要。结论 RoCC隐蔽通信框架综合可证明安全隐写、生成式机器学习方法和跨模态检索方法的优势,与现有的方法比较,具有更加隐蔽和安全的优势,并且是当前对数据传输丢包异常最鲁棒的模型。Objective Covert communication is a pivotal research area in the field of information security.A highly covertand secure covert channel for transmitting sensitive information must be developed to safeguard the privacy of communica⁃tion users and prevent occurrences of eavesdropping on confidential data transmissions.Most methods build covert chan⁃nels by tunneling multimedia streams.However,the problem of packet loss caused by fluctuations in network transmissionis not considered.This study proposes a covert communication method that is robust to network anomalies and is based oncross-modal information retrieval and provably secure steganography.Method We propose a general covert communicationframework named robust covert communication(RoCC),which is based on cross-modal information retrieval and provablysecure steganography.Artificially generated information from artificial intelligence(AI)systems,including deep synthesismodels,AI-driven artwork,intelligent voice assistants,and conversational chatbots,has emerged.These AI models can synthesize multimodal data,such as videos,images,audio,and text.The practical application of provably secure steg⁃anography has become a reality as generative models make significant strides.Thus,we introduce generative models andprovably secure steganography techniques into our framework,embedding secret messages within the cover text data.Fur⁃thermore,the domain of speech synthesis and recognition has witnessed the advent of numerous mature open-source mod⁃els,facilitating seamless cross-modal conversion between speech and text.Our approach employs a combination of directand indirect communication.In direct communication using voice over internet protocol(VoIP)network call service,realtime synthesized audio stream data are delivered,and the receiver can restore the text through voice recognition.Indirectcommunication uses a public network database for steganographic text data transmission.The receiver restores lost textsemantics because of network packet loss and speech r

关 键 词：信息隐藏 隐蔽通信 生成式模型 数据跨模态转换 可证明安全隐写 多媒体信息检索 相似度分析 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]