Pierre Moulin U of Illinois @ Urbana Champaign
A variety of technologies have recently become popular for embedding one set of data into another; this includes watermarking, steganography, fingerprinting and traitor tracing. These technologies are finding applications in emerging areas such as intellectual property protection, media authentication, database indexing, broadcast encryption, broadcast monitoring, and covert communications. The application areas are expanding: they also include communication problems such as broadcast channels and wireless communications. While the development of these technologies has largely been heuristic, recent research has demonstrated the feasibility of a more scientific approach rooted in information theory. This project expands this scientific framework and explores the multiple fundamental connections between the signal embedding applications listed above. One central theme is the role of binning schemes for information transmission; random binning schemes are known to be capacity-achieving in problems of communications with side information at the encoder and/or decoder. Yet these analyses are based on restrictive assumptions (such as coherent communication channels) which have limited the practical applicability of binning schemes so far. The second central theme of this project is the premise that security requirements impose constraints which are naturally addressed by a complexity-based cryptographic approach rather than a purely information-theoretic or decision-theoretic approach. This project develops a systematic framework for integrating the information-theoretic and the cryptographic aspects of secure signal embedding problems. This project is highly multidisciplinary and involves a synergy between research and educational activities in signal processing, communications, coding theory, information theory, and cryptography.
