有限先验情景下的隐写分析技术研究

Research on Stegnanlytic Techniques Under the Condition of Limited Prior Knowledge

作为隐写技术的对抗手段,隐写分析主要致力于检测、阻止甚至破坏互联网上可能存在的基于隐写的隐蔽通信行为.实用隐写分析技术的研究,对打击恐怖组织及不法分子对隐写技术的滥用、维护国家安全和社会稳定具有重要的意义.首先阐述了隐写分析技术的研究意义及研究现状,然后,从需求极为迫切的图像隐写分析切入,针对有限先验信息的情境,提出新的隐写分析研究框架.以海量多样代表性样本数据为研究对象,重点研究具有超完备性的隐写分析特征基构造方法,利用超完备特征基对测试样本进行表示,根据表示系数对隐写方法进行预测,为研制准确性更高、实用性更强的隐写分析方法提供理论和技术支撑.

Steganalysis, as the countermeasure of steganography, mainly concentrates on the detection , prevention and destruction of the possible steganography based on covert communication behaviors on the Internet. Research in practical steganographic analysis technology plays an important role in fighting against the steganography abuse by terrorist groups and criminals and maintaining national security and social stability. This paper firstly introduces the research significance and related works in the past, and then proposes the new steganalysis framework to address urgently needed image steganalysis with limited prior knowledge. Taking plentiful, multiple and typical data as study objects, this paper focuses on studying the methods to construct over-completed steganographic analysis feature bases . using these feature bases to represent testing samples and predicting the steganography method of testing samples according to the representing coefficients, which provides theoretical and technical support for developing steganalysis schemes with higher accuracy and better practicality.