Author(s)

V. Nosso, F. Garzia, R. Cusani

Abstract

Iris recognition is regarded as the most reliable and accurate biometric identification system available. The present work involves the development of a novel technique in order to improve the performance of iris recognition systems. We have used for our experiments a publicly available iris recognition system. Tests on CASIAv3 image database have resulted in a 2% accuracy improvement with respect to traditional methods; a significant one in iris recognition.

Keywords

iris recognition, PCA, genetic algorithms, cosine distance, biometrics