A FUZZY DNA-based Algorithm for identification and authentication in an Iris Detection System

Abstract

Several algorithms have been proposed in order to identify or authenticate a certain user in an iris detection system. These previous algorithms are based on different transforms applied to a certain features extracted from the iris, known as iris template. However, this new approach aims to identify and/or authenticate a user avoiding transforms, comparing directly the Iris templates instead. Based on ideas inherited from DNA alignment algorithms and fuzzy logic, this new algorithm attains a better result in deciding whether two templates belong to the same user, when being compared to previous algorithms. Furthermore, this new algorithm allows a partial comparison between the whole template stored in a database, and a portion of template, decreasing the accessing time. As a result, not only false acceptance rate (FAR) and false rejection rate (FRR) but also time in identifying and authenticating are highly decreased compared to current algorithms and furthermore its flexibility in a local comparison makes of it a promising algorithm in not only an iris detection system, but in any biometric system.