Adaptation of integrodifferential operator using statistical characteristics of the iris texture

Recognition of humans based on characteristic eye features has taken a significant place in security and identification systems in the past two decades. The complex pattern of the iris, that is unique to an eye, makes iris a great biometric descriptor. The first step in the process of the biometric identification based on iris is its localization in an image. The shape of the iris, its size, position, shades from the light sources and eyelashes make this process difficult. It is reported that most failures to match in iris recognition system result from inaccurate segmentation. This paper will analyze statistical characteristics of iris texture with an attempt to improve the process of segmentation.