2265 Beckman Institute
405 N. Mathews
Urbana, IL 61801
Pierre Moulin received his Engineer degree from the Ecole Polytechnique of Mons, Belgium, and his doctorate from Washington University in St. Louis (1990). After working as a Research Scientist for Bell Communications Research in Morristown, New Jersey, he joined the University of Illinois as Assistant Professor (1996) and later became Associate Professor (1999) and Professor (2003) in the Department of Electrical and Computer Engineering, Research Professor in the Coordinated Science Laboratory, faculty member in the Beckman Institute's Image Formation and Processing Group, and affiliate professor in the department of Statistics. He is also a member of the Information Trust Institute and the founding director of the new Center for Information Forensics, a multidisciplinary research center currently involving twenty colleagues. His fields of professional interest are information theory, image and video processing, statistical signal processing and modeling, decision theory, information hiding and authentication, and the application of multiresolution signal analysis, optimization theory, and fast algorithms to these areas.
He served as Associate Editor for the IEEE Transactions on Information Theory from 1996 till 1998, for the IEEE Transactions on Image Processing from 1999 till 2002, and then as Area Editor from 2002 till 2006. In 1999, he was co-chair of the IEEE Information Theory workshop on Detection, Estimation and Classification. He was a Guest Editor of the IEEE Transactions on Information Theory 2000 special issue on Information-Theoretic Imaging; Guest Editor of the IEEE Transactions on Signal Processing's 2003 special issue on Data Hiding; and member of the IEEE Image and Multidimensional Signal Processing (IMDSP) Society Technical Committee (1998-2003) and the Board of Governors of the IEEE Signal Processing Society (2005-2007). He is co-founder and Editor-in-Chief of the new IEEE Transactions on Information Forensics and Security. He is a Fellow of IEEE (2003), recipient of a 1997 Career award from the National Science Foundation, and of the IEEE Signal Processing Society 1997 Best Paper award in the IMDSP area. He is also co-author (with Juan Liu) of a paper that received the IEEE Signal Processing Society 2002 Young Author Best Paper award in the IMDSP area. He was selected as Beckman Associate of UIUC's Center for Advanced Study (2003) and Sony Faculty Scholar (2005-2007). He was on the Dean's list of teachers rated excellent by their students in 1996, 1999, 2000, 2005, and 2007.
KEY WORDS: Signal processing, modeling, compression, imaging, information theory, statistics, stochastic processes, wavelets and multiresolution analysis, estimation, detection, pattern recognition, watermarking, hashing, fingerprinting, steganography, information security, optimization.
Moulin and his students research the application of statistics and information theory to a variety of engineering problems. Generally speaking, the main challenge is to select emerging problems for which a fundamental understanding is still lacking, identify the underlying simple and elegant principles that should be brought to bear on such problems, and develop methods and algorithms for solving them. Statistical modeling and fundamental principles of statistical inference and information theory play a central role in this approach. The application areas include imaging and signal, image and video processing, with focus on problems of compression, restoration, detection, and more recently, information forensics and security.
One of our research thrusts includes finding sparse signal representations (e.g., wavelet-based) and processing techniques that optimize appropriate performance measures. This line of research aims at broadening the scope of information-theoretic methods and results on optimal signal processing, which currently apply to somewhat narrow classes of idealized signals. Examples of this work include design of wavelets and filter banks optimally adapted to signal statistics; optimal quantization techniques for nonorthogonal filter banks; modeling, estimation and coding of dense video motion fields; spectral density estimation, radar imaging, complexity-regularized image restoration, and Bayesian image restoration.
Our main research thrust nowadays is in the emerging area of information forensics and security. Moulin's interest in information security dates back from 1997, when he developed an information theory of watermarking and data hiding, together with Prof. Jody O' Sullivan (Washington U.). This theory finds applications to content protection; previously, only heuristic methods were available. Since 2000, we have been developing a coding theory for data hiding, together with Prof. Ralf Koetter (formerly UIUC, now TU Munich). Examples of this work include fundamental performance bounds for watermarking, digital fingerprinting, and steganographic systems; codes that can achieve (or approach) these bounds; methods that resist arbitrary noise and desynchronization attacks; robust hashing methods; analysis of security flaws for watermarking systems; methods that can remedy those flaws; and information-theoretic analysis of secure communication systems (joint work with Prof. Muriel Medard, MIT).