Iris is rapidly gaining acceptance and support as a viable biometric. Several large scale identity management applications are either using or considering iris as their secondary or primary biometric for verification. While there are several academic publications addressing the problem of iris image quality, NIST Iris Quality Calibration and Evaluation (IQCE) is the first public challenge in iris image quality aimed at identifying iris image quality components that are algorithm- or camera-agnostic. IQCE evaluated 14 iris image quality assessment algorithms in their effectiveness in predicting the performance of iris recognition algorithms, their computational efficiency and their robustness. Interestingly, the implementations which are the best predictor of recognition performance are also the fastset (i.e., the shortest computation time) and with zero failure to computation, are the most robust. To quantitatively support the development of the international iris image quality standard (ISO/IEC 29794-6), IQCE also examined the effect of 14 iris image covariates on performance. This evaluation supports homeland security, counter-terrorism, and border control applications by enhancing reliability and accuracy of iris recognition, and significantly improves requirement planning and system design.
Citation: NIST Interagency/Internal Report (NISTIR) - 7820
NIST Pub Series: NIST Interagency/Internal Report (NISTIR)
Pub Type: NIST Pubs
biometrics, biometric quality, iris, recognition error