|
The project focuses on advancing the field of digital image steganography -- a covert way of communication in which information is hidden in other objects, such as digital media files, to assure privacy. For a secure steganographic system, it should be impossible to prove the presence of hidden data. Achieving this level of security in practice is extraordinarily difficult because digital media is hard to describe using statistical models with accuracy necessary to guarantee perfect security.
This project works with devices that acquire an image to learn a sufficiently accurate model within which it becomes possible to construct steganography with a verifiable level of security and performance that can be contrasted with theoretically achievable limits. The result is a novel advanced privacy tool needed in countries that prohibit the use of encryption or in a hostile environment when there is need to communicate without attracting attention via channels in control of an adversary or via public, insecure channels. On the other hand, a deeper understanding of the limits of covert communication will facilitate better defense against such methods of deception, an effort recognized as steganalysis. The project explores commercial applications of data hiding include signal authentication, integrity verification, and secure data dissemination. This research contributes to trusted information exchange, data mining, information assurance, network and computer security, counter-deception, and intrusion detection and its prevention.
|
TWC: Medium: Steganography in Empirical Sources Using Acquisition Oracle