Iris segmentation based on Fuzzy Mathematical Morphology, Neural Networks and ontologies

Abstract

Segmentation is one of the most time-consuming steps within the whole process of Iris Recognition. By means of Fuzzy Mathematical Morphology and Neural Networks, this new algorithm can fulfill the task of isolating the Iris, not only with an acceptable accuracy, but also with a very high improvement in terms of time. Furthermore, this innovative scheme presents an ontology able to decide whether the features can be extracted, based on previous segmentation. This paper provides a detailed explanation of both the problem to be solved and how this new approach meets the required goals. Current Iris Recognition algorithms may benefit from this new approach, and what is more, the essence of the algorithm can be extended to other biometric segmentation procedures.