By discovering and quantifying the “limbus effect,” Oak Ridge National Laboratory researchers have advanced the state of the art for human iris recognition systems. While the iris is a proven and reliable biometric for verification or identification, non-ideal images – such as those captured off axis – are problematic. Reasons include cornea refraction and the limbus effect, causing iris recognition performance to decrease and in many cases fail entirely. Now, using an anatomically accurate human eye model and some slick math, a team led by Chris Boehnen of ORNL and Ed Chaum, an ophthalmologist at the University of Tennessee-Memphis, has demonstrated a technique to improve off-axis iris recognition.