Input images in iris recognition systems: A case study

Iris pattern is a unique, stable and non-invasive biometric feature, suitable for individual recognition purposes. There are several and very diverse iris recognition algorithms, but in most cases, a collaborative environment and ideal conditions are required when capturing the system input image. To overcome these constraints and increase the number of domains in which iris recognition systems can be used, it is essential to develop robust algorithms that work in non-collaborative environments, but in order to do this in a correct way, it is important to previously analyse the different situations that can occur in such environments. In this context, different noisy and artificial iris images are analysed in this paper in order to determine its influence in iris recognition systems performance.