VoIP流媒体隐写的在线检测模型与方法研究

Steganography and steganalysis technologies based on VoIP which take bit stream as the carries are being becoming researching hot-spot. Different from the traditional static documentary carries through images and videos, VoIP is characterized of multimode, real-time and dynamic, which challenges traditional steganalysis technology . The main purpose of this project is to explore the correspondent steganalysis model and methods. The research contents include two aspects as following: the establishment of a unified online detection model for VoIP multimode steganography and the correspondent steganalysis methods against new steganography approaches targeted at VoIP stream, including QIM steganalysis in low rate speech, low embeding rate LSB steganalysis in the high bit-rate PCM speech codec and RTP/RTCP protocol. According to the basis, a online detection system of compatible steganalysis algorithms and good extendibility of VoIP streaming media steganography are developed. The establishment of model and key technology breakthrough of steganalysis methods will setup a theoretical foundation for the practical application of the corresponding steganalysis system, development of steganalysis application system provides the technical support for the national security, social stability in china.

以VoIP（Voice over IP）流为载体的信息隐写及隐写检测技术是当前的研究热点。与传统图像、视频等静态文件载体不同，VoIP流媒体信息隐写具有多模式、实时性和动态性特点，这给传统的隐写检测技术的应用带来了挑战。本项目拟以探索与VoIP流媒体信息隐写相适应的隐写检测模型和检测方法为研究目标，进行两个方面研究：建立VoIP流多模式信息隐写在线检测模型；针对VoIP流最新隐写方法研究相对应的隐写检测方法，包括低速率语音编码中QIM隐写检测，高速率PCM语音编码中小嵌入率LSB隐写检测，以及RTP/RTCP协议的隐写检测等。在此基础上研发一套兼容多种隐写检测算法和具有较好扩展能力的VoIP流信息隐写在线检测系统原型。在线检测理论模型的建立和检测方法关键技术的突破将为隐写检测系统的实际应用奠定理论和技术基础，而应用系统的研发为我国的国防安全、社会稳定能提供技术保障。

信息隐写相比于传统的加密技术，它不仅隐藏了隐秘信息的内容(不可读性)，而且掩盖了隐秘信息存在的事实(不可见性)，安全性更高，对国家安全、经济发展和社会稳定均有影响。所以流媒体为载体的信息隐写及隐写检测技术日益受到重视。本项目以VoIP多模式信息隐写在线检测模型和检测方法为研究对象，解决了VoIP多模式信息隐写的在线和统一实时检测问题。实现了一套 VoIP 隐写检测的实验原型系统，验证了理论成果的有效性。完成的研究内容有：VoIP多模式信息隐写在线检测模型、VoIP流量的快速准确捕获和信息隐写检测等关键技术等。本项目发表论文11篇，其中SCI论文6篇，会议论文2篇（EI会议1篇），中文核心3篇；申请技术发明专利2项；获得国内学术会议优秀论文1篇；专著1部。本项目的研究成果极大地推动信息隐写技术应用的基础研究。