Large scale iris image quality evaluation

Abstract

Several recent studies have shown that while iris images captured at near infrared are viable biometrics for verification and identification, similar to other biometrics, its performance drops when comparing images from imperfect sources (e.g. subject blinking), under imperfect conditions (e.g. out of focus) or non-ideal capture device. The immediate question to ask is what factors and to what degree are most influential on iris recognition performance. Motivated by this need, National Institute of Standards and Technology (NIST) initiated Iris Quality Evaluation and Calibration (IQCE). IQCE aims to define and quantify iris image properties that are influential on performance of iris recognition. This paper gives an overview of the IQCE.