Would you please describe the Center’s goals?
The Center for Identification Technology Research (CITeR) is a National Science Foundation sponsored, so-called, “industry university co-operative research center”. Our mission is to advance identification technology, and the research is centered around biometric systems and credibility assessment. We provide research in the cross-cutting and enabling disciplines and technologies related to applications of biometrics, human credibility assessment and in general, the issues that are related to identity management and the meaning of identity in a global society.
We are accomplishing this through the collaborative research that is done between different universities along with the partnerships that are in place between companies and government agencies. In the process, we also train interdisciplinary scientists, engineers, students and manage to facilitate technology transfer through the creation of excellently educated workforce that eventually finds employment within either the private sector or within the government. In such a way, we have an entire cycle of activity that has been working for going on ten years now.
How is the Center funded?
The Center is funded through its affiliates. Allow me to explain what I mean. Industry university co-operative research center concept has been developed by the National Science Foundation in the early 1980’s. The program provides structure for companies and government agencies to provide funding for and collaborate with university researchers. And, for that funding, they fundamentally choose what kinds of research the university will be performing. More specifically, our Center has approximately 20 to 25 affiliates, and these affiliates include the largest national and international system integrators – companies such as Accenture, Booz Allen Hamilton, Northrop Grumman, Lockheed Martin, Computer Science Corporation, MorphoTrak, SAIC, Raytheon and others, as well as about a dozen federal agencies. These companies and agencies pay an annual affiliation fee, which, at this point in time, is 40 thousand dollars. These affiliation fees then create a pool of dollars which directly fund the research being done at the universities. Along with West Virginia University, we have the participation from University of Arizona in Tucson, Clarkson University will also be joining the Center, and we also collaborate with the Michigan State University.
University researchers talk to representatives from these agencies and companies and develop project proposals. These proposals are then critically evaluated by the affiliates – the industry and government representatives. Proposals that best meet the needs and requirements are then selected for funding. Every year, the Center runs about ten to fifteen projects and, as I said, these projects are selected and monitored by the affiliates.
Excellent; and could you tell us, please, what are some of the key projects that the Center is currently involved with?
Certainly; the Center was created in 2001, and since then, we have run about 80 to 90 projects. This is, obviously a large number. At the same time, I must inform you that these projects are all one year long; in cases when generated results are particularly interesting, the project can be extended to two years. Some of the projects that have been selected for funding at our last meetings, which occurred this past May, include the study of mid-wave infrared light images for facial recognition. We attempt to evaluate how facial recognition could be performed using images captured across different light bands. What is especially interesting, are the aspects of “liveness” features that you can add to facial recognition using infrared illumination. We are also actively working on cross-age facial recognition projects, and our more recent research has been heading in the direction of utilizing DNA as biometrics. In this case we are considering highly degraded samples, for example, and avoid a long chemical process.
We’re also keenly interested in what it means to design effective and efficient large scale biometric system. For instance, we modelled systems similar to US Visit, and from there we are learning about the risks imposed on system performance by either the false accept or false reject rate of biometric systems how do they affect overall system security, how do they affect the length of the lines at check points, how do they affect the computational infrastructure and data base searches within the system?
We always have interesting projects in known modalities; iris and ocular biometrics, face and fingerprints. Recent projects deal with an understanding of altered fingerprints and how can we catch up with that threat to security. On the other side, we have some very vigorous projects that are trying to establish the credibility of speakers in any video or audio file. These are, as I would describe it, non-intrusive lie detection tests.
We also have research that tackles problems related to surveillance, the development of low cost camera sensors and sensor networks and optimization with respect to performing biometric identification from such networks.
Impressive, you certainly have many projects on-the-go.
Indeed; in fact, at any point in time, the Center has, as indicated, about ten to fifteen projects actively being worked on. Of course, we also have a number of investigators involved. Faculty members and students with very broad and extensive backgrounds participate in the projects on identity management. The backgrounds of our investigators range from computer security and software engineering to pattern recognition, computer vision, machine learning. Our researchers are affiliated with engineering colleges, business schools and communication departments, because of the link between credibility assessments and the use of verbal and non verbal communication. As a result, we are a truly diverse Center which represents a broad spectrum of qualified investigators. This, I believe, is our strength as it prevents us from viewing the problems that can be found in biometrics simplistically. Due to our diversity, we are capable of bringing expertise to viewing the system as part of the computational infrastructure; this then means we are able to analyze the security implications, the interfacing between the systems and their specific users, the privacy implications along with the computational infrastructure that is required in order to support large scale identification and identity management systems. It is really a much broader focus than “just biometrics”.
I also understand that you’re involved with the FBI’s Biometrics Center of Excellence; could you please tell us a little about this too?
Certainly. As you are, undoubtedly, aware FBI has a tremendous experience in developing automated identification systems. The agency has been an affiliate of CITeR since its inception in 2001/02. Based on existing collaborations, West Virginia University became FBI’s academic partner for biometrics research and plays the role of the portal for the academic research community. This role involves active partnering with universities and research labs across the US. The research topics are unique with respect to the applications of interest to the agency. The projects sponsored by the FBI include large scale biometric data collections as well as the analysis of performance barriers and the trust one can place upon biometric systems. Identification systems are, undoubtedly, going to continue to expand in terms of the size of repositories, complexity, in the number of modalities, and even in the flexibility of their use. Understanding these changes implies that we can analyze their implications and counter operational challenges within the next five, ten or fifteen years.
I hear you mentioning looking down the road some five to ten years for some of this information; on that note, where do you see biometrics being used for the average person ten years down the road, for example?
I must admit, that is a somewhat difficult question. Let me say that I believe that the future of biometrics will be following the future of computation, as we know it today. I recall when I was in high school and then attending university, we were all very excited by mainframes, and were truly enthusiastic about being able to use personal computers and so on. Currently, computing platforms are pretty much everywhere; cameras are almost in every device these days, we have wearable devices.
As a result, it’s my view that biometrics represents the natural progression, or choice, if you will, for authentication; human authentication, in the future, means we’re going to be mutually authenticated. Devices will be authenticating ourselves, we will be authenticating any devices which we’ll be using and many of those authentication tasks will be performed by way of biometrics. This doesn’t mean, necessarily, that we will be done with passwords altogether. We must be cognizant of privacy issues and possible violations caused by the naive use of biometrics. We will have to ensure that what we choose to share on the network are ”meaningless” binary representations which are impossible to convert back into the feature sets which represent individuals. So of course, security aspects are going to be highly important.
I honestly believe that biometrics is going to be pervasive, all around us, and we won’t even recognize it as something special or out of the ordinary.
You mention privacy issues and how one must be watchful of how biometrics affects one’s privacy; would you then say that this aspect is one of biometrics’ greatest challenges, and/or are there other challenges that you foresee for the future?
I would agree that dealing with any ethical issues which surround biometrics and, naturally, privacy policies, are some of the challenges. However, I personally believe that we already have the technical capabilities which address, to a large extent, privacy concerns. The issue really is, are we, as a society, and as the people who are both deploying and paying for the deployment of biometrics, are we going to ensure that the best privacy practices are deployed within each and every system? I think that every vendor that I’m aware of, and with whom I’ve spoken, is very aware that they ought not allow privacy violations to take place in any biometric application. I truly believe that there is an excellent level of awareness in the industry as a whole. There is certainly a great level of awareness in the government sector, and within academia. Possible violations in terms of the unethical use of biometrics and privacy are, actually, the threat for the economic prosperity of the biometric industry. Everyone is being especially careful to ensure they employ state of the art techniques that provide the adequate level of protection.
As far as the second portion of your question, regarding other challenges that are facing biometrics, I would have to say that over the past five years, we have noticed what seems to be a very interesting trend. Firstly, we have seen the end of small biometric system and device providers. Along with a consolidation within the industry, many companies today offer biometric services and products and have the capability to deal with system integration and understand the context in which biometric systems are used. They also deal with security solutions, not just biometric solution. I think this trend in the industry is also reflected in the way government, industry and universities work together. We cannot have university research being separated from the private sector and the government sector because each of them brings specific and unique strengths to the table. So, I think that’s been a trend that I have personally observed, the unification of capabilities and increased cooperation.
At the same time, with respect to technical challenges, I think there are many. Biometric systems, at this point in time, require rather cooperative users. Biometric systems also require reasonably good environmental conditions, in terms of lighting, the proximity of cameras or sensors, and even in terms of the assumptions the user is not intentionally trying to fool the system. I see that all of those assumptions will have to be relaxed in the systems that strive to be marketable in the near future.
With regards to the future, we know we have significant capabilities in face, iris, fingerprint, vein pattern recognition; all the modalities that we know of as biometrics today are very well developed. Now we need to understand what happens when people try to fool these devices. We’re talking now about liveness detection and reducing the vulnerabilities since biometrics, fundamentally, is part of security protection mechanism of the system.
We will also be seeing the deployment of biometrics in uncontrolled environments. We need to develop systems which will be able to use biometrics remotely, where there is no one monitoring the sensor being used to acquire the physiological or behavioural signature of an individual. All of these issues represent very significant challenges that are currently in the fundamental and to some degree, applied research. We will be seeing products and solutions come to market within the timeframe that you mentioned; some within the next two years, other will take five to ten years.
In summary, the model of cooperation that CITeR offers to affiliates has been demonstrated to be
successful and I encourage you to talk to some of the industry and government representatives in our Center in order to get their opinion on what the Center is doing and how they are benefiting from it. At the same time, I want to emphasize that CITeR is not a “closed society”. We are always open to new affiliates, to companies that believe that they can benefit from the types of work being performed at the Center, that can gain from a number of students that are being educated at WVU and other universities with degrees in biometrics and related disciplines. We would be thrilled to talk to industry representatives, if there is an interest in participating with the Center.
That sounds wonderful; and I want to thank you very much for filling us in on all the activities at CITeR; it certainly sounds as though you have some wonderful projects underway. Congratulations on the work that you’ve already accomplished within the past nine years. Thanks again for your time with us.
You are most welcome Peter.
CITeR is a National Science Foundation (NSF) Industry/University Cooperative Research Center (I/UCRC). Its mission to advance identification technology is strongly focused in the areas of biometric systems and credibility assessment. Our portfolio of activities achieve this mission through cross-cutting research of emerging enabling technologies, interdisciplinary training of scientists and engineers, and facilitation of technology transfer to the private and government sectors through its affiliates.
West Virginia University is the Center’s founding and lead I/UCRC site focusing on biometrics and related identification technology and systems. The University of Arizona is the Center’s second site focusing on credibility assessment systems. Each CITeR University site maintains interdisciplinary collaborative partnerships with other academic institutions to effectively respond to affiliate research needs.
Reproduced please specify：zkblog » Interview with Bojan Cukic, Co-director, The Center for Identification Technology Research