Czajka, A; Piech, K
This paper describes an application of the Zak-Gabor-based iris coding to build a secure biometric verification station (SBS), consisting of a professional iris capture camera, a processing unit with specially designed iris recognition and communication software, as well as a display (LCD). Specially designed protocol controls the access to the station and secures the communication between the station and the external world. Reliability of the Zak-Gabor-based coding, similarly to other wavelet-based methods, strongly depends on appropriate choice of the wavelets employed in image coding. This choice cannot be arbitrary and should be adequate to the employed iris image quality. Thus in this paper we propose an automatic iris feature selection mechanism employing, among others, the minimum redundancy, maximum relevance (mRMR) methodology as one, yet important, step to assess the optimal set of wavelets used in this iris recognition application. System reliability is assessed with approximately 1000 iris images collected by the station for 50 different eyes.