Indian Journal of Science and Technology

Abstract

Background/Objectives: One of the most robust biometric based recognitions in existence is iris recognition, which has many practical applications in many fields including live passwords and forensics etc. Governments use iris codes as unique code to identify an individual (used in AADHAR). Even if iris recognition system is considered as impossible to break and it is highly secured, it is prone to some vulnerability. Attacks like indirect and direct attacks can be performed in iris based recognition for penetrating in to the system and fraudulently accessing it. In this paper, a method is proposed, which will determine a user as genuine or an imposter. Methods/Statistical Analysis: For detecting whether an imposter wearing a textured lens, BoW (Bag of Words) method is used. Findings: In this paper the biometric template created from the iris which does not have contact lens is matched against the previously stored biometric templates. Confusion matrix for lens detection is calculated. Images from two databases are used for experimental purposes, IIIT-D contact lens database and CASIA-Iris-interval database. Application/Improvement: BoW method improves the performance of the existing method for contact lens detection.

Keywords: Bag of Words, Indirect attack, Iris recognition, Vulnerability