(Story and photo provided by Augusta University)
As cyber attackers continue to find new and advanced ways to circumvent facial recognition and fingerprints, Sejun Song, a professor of computer science in Augusta University’s School of Computer and Cyber Sciences, believes “EyeDentity” will offer a new perspective on identity verification.
Earlier this year, the National Security Innovation Network hosted the Innovators Challenge where representatives from top schools in the United States presented groundbreaking solutions to the Department of Defense. The competition was divided into three categories – modeling and predictive analytics, persona and influence, and data and permeability – with teams vying for top prizes of $150,000, $100,000 and $75,000.
Song clinched third place in the persona and influence category with his innovative, solutions-based research, “EyeDentity: Eye Reflections for Identity Authentication against Liveness and DeepFake Attacks.”
The category focused on understanding how attackers disguise themselves and create fake online identities to avoid getting caught. Teams were tasked with finding weaknesses and uncovering fake identities to improve the accuracy of identifying and tracking them.
“Face recognition-based touchless identity authentication applications, such as those in unmanned kiosks, smartphone security, airport check-ins and banks, are prevalent; however, concerns surrounding privacy violations, fake presentations, potential biases and security vulnerabilities have arisen with the widespread adoption of face recognition systems,” Song said. “’EyeDentity’ doesn’t just look at the shape of the reflections but also considers details like the environment, color, and light, making the system more accurate and reliable.”
Unlike facial or fingerprint recognition, Song’s work stands out by using a unique method to enhance facial recognition through the analysis of reflections in an individual’s eyes. It operates by displaying various shapes and colors on the screen while the camera captures how these images are reflected in the eyes. The system then employs AI to analyze these reflections and determine whether a person is real or a deepfake.
The idea for “EyeDentity” was inspired by a touching moment with Song’s daughter who said, “I see me in your eyes!” This simple observation led to the insight that “your eyes show what your eyes see,” said Song.
Song emphasized the importance of prioritizing liveness over secrecy in biometric systems and keeping solutions straightforward.
Song hopes “EyeDentity” will become a smooth and effortless authentication system that works quietly in the background, transforming authentication and boosting security for everyone by eliminating the need for passwords.
Augusta Good News is an award-winning member of the Georgia Press Association. Sign up for our weekly newsletter here.