As you watch a crime scene investigator in a movie look for evidence, know that the movie itself is providing a clue toward crime prevention. Most movies today include an on-screen yet invisible anti-piracy digital fingerprint that stamps the individual theater showing the motion picture – a fact which ties in to Beckman Institute faculty member Pierre Moulin’s research into the forensics of video and images.
Moulin, a Professor in the Department of Electrical and Computer Engineering and a full-time faculty member in Beckman’s Image Formation and Processing Group, is a leader in decoding those clues through his research into information hiding and extraction in images and video. He is the director of the Center for Information Forensics at the University of Illinois’ Information Trust Institute, where he studies issues such as digital fingerprinting.
“It actually is used a lot in movie theaters,” Moulin says of the technique of digital fingerprinting. “You can’t see it, you just see the movie and you don’t even know it has a fingerprint. It is projected on the screen, you just don’t see it.”
Moulin said modern movie piracy often involves recording with a high-quality camcorder in a movie theater that has been pre-arranged to include only the one doing the filming and the projectionist. A recent study estimates film piracy costs the industry more than $6B a year, with more than 80 percent of the piracy taking place overseas.
Moulin describes a digital fingerprint as “an invisible pattern superimposed to the image (in case of image fingerprinting) or an inaudible pattern in the case of audio.” He said that once a pirated DVD is obtained, then the question becomes one of how to extract the digital fingerprint from it.
“These techniques are being used right now and in fact they catch people this way,” he said.
And that is where Moulin’s work comes in. He has developed applications for extracting digital fingerprints that are currently in use worldwide.
“My contribution in the field has been, what are the limits of what can be done, and what kind of fingerprinting technique can be developed which can approach those limits?” Moulin said.
Moulin, who earned an engineering degree in his native Belgium and a Ph.D. in electrical engineering from Washington University in St. Louis, came to the University of Illinois and Beckman in 1996 from private industry, where he worked for Bell Communications Research.
Moulin has a research focus on topics such as image and video processing, information hiding and authentication, statistical signal processing and modeling, and the application of multi-resolution signal analysis, optimization theory, and fast algorithms to the areas he is researching. He said mathematical theories undergird his work.
“My research in the field started with developing a mathematical theory for those problems. It’s a pretty big field so it took a number of years,” he said. “Based on the mathematical theory, then one can construct algorithms or methods which can be shown to be nearly optimal, where it’s almost impossible to improve them. There is only so much you can hide, with the requirements that it should be invisible and also detectable.”
Hiding information such as digital fingerprints in a movie and extracting that information when the need arises is the key focus of work at the Center for Information Forensics.
“It is a field that is quite multidisciplinary because it addresses issues that pertain to content of images and video and how you present them efficiently, and also to coding techniques and information theory,” Moulin said. “There are several fields that are essential to doing this kind of work. The idea for a center like this is to put people with different expertise together so that it makes it easier to develop a solution.”
Moulin knows that research doesn’t always lead to solutions and commercial applications, especially in a field like image and video processing, but that goal is what drives his research.
“When I select a research topic, it should have an application,” he said. “That’s always a risk in this field. There are many techniques that were promising commercially and for some reason they were never applied. But that’s part of research – you never know what will be the commercial or economic impact.”
Doing research with an eye toward future applications has always been a credo of efforts at the Beckman institute.
“It’s a wonderful place with great colleagues and a great environment,” Moulin said of working at Beckman. “What makes me happy is research that is mathematically interesting and challenging. At the same time it should be useful. I spend a lot of time trying to select my topics because I want to choose ones that would be useful.”
This article is part of the Winter 2010 Synergy Issue, a publication of the Communications Office of the Beckman Institute.