Everyone is unique.
This is the basic premise upon which biometric authentication is founded. Biometrics, the science of measuring and analyzing human physical and behavioral characteristics, is not itself a new field of study. However, in the 19th and early 20th centuries biometric authentication began to emerge as its own field of study as scientists realized that certain human physical characteristics, such as fingerprints, were unique to each person and could be used for identification purposes. Fingerprinting is perhaps the most widely recognized form of biometric authentication, having been popularized in the cultural imagination through a variety of movies and TV shows depicting scenes of criminal investigations or bad guys being booked at the local police station.
Other forms of biometric authentication have become increasingly common over the last few decades and are currently being used in a diverse number of applications including corporate and public security systems, military surveillance, counter-terrorism initiatives, and point of sale software. Now, a group of researchers at the CyLab Biometrics Center is pushing the envelope by merging the time-tested science of biometrics with the burgeoning field of information technology. Under the direction of Dr. Marios Savvides, the CyLab Biometrics Center is harnessing the power of computing to advance the state of the art in biometric authentication technologies and meet an array of new, 21st century security challenges.
Much of the work at the Biometrics Center is focused in the areas of iris and face recognition. Research in iris recognition primarily addresses how to improve image acquisition and image quality. Savvides and his team are exploring ways to make iris recognition devices, such as the Iris-on-the-Move ™ (IOM) system developed by Sarnoff Corporation, operate more effectively by improving their ability to capture usable iris images from people moving in a crowd at a distance of approximately 10 feet. Additional research in iris recognition is investigating ways to improve image processing for enhanced quality through the use of more efficient, robust segmentation algorithms. Algorithms that work more efficiently can significantly reduce the amount of time it takes for a scanning device to segment an iris image, which in turn can yield fewer instances of images that include flaws such as specular reflections.
Researchers at the Biometrics Center are also making great strides in developing better technologies for use in face recognition and detection systems. Two of the biggest challenges being addressed are image capturing and pose correction. Currently, surveillance cameras have an image capture rate of only 60 frames per second. Savvides and his team are developing a new algorithm which nearly doubles that rate to 100 frames per second, thereby increasing the likelihood of detecting those fractions of a second that may yield a usable face image. Further research in the area of face recognition addresses problems of pose correction, with projects currently underway on a new process to correct a non-frontal image by fitting a 3D morphable model onto a 2D image, then rotating the model into a frontal 2D image suitable for enrollment in face recognition databases.
Biometrically-based security systems are already being used for military surveillance, fighting terrorism, and improving national security. As more and more systems are deployed to ensure security and increase the safety of our nation, the biometric authentication technologies being developed at the CyLab Biometrics Center will be critical to their success and their commercial implications will be many, according to Savvides.
The Biometrics Center is just one several research centers at CyLab leading the way in the development of real-world cyber security solutions. To learn more, please visit the Center website.