New methodology for a person identification system

Abstract

Reliable person identification is a key factor for any safety measure. Unlike other biometrics such as the palm, retina, gait, face and fingerprints, the characteristic of the iris is stable in a person's lifetime. Iris patterns are chaotically distributed and well suited for recognizing persons throughout their lifetime with a single conscription. This paper proposes a new approach to person recognition based on iris patterns, which works with indoor outdoor conditions, spectacles contact lens wearing persons and diseased eyes. A challenge-response method is used for eye aliveness checking that puts off artificial sources from entering the iris database. The proposed algorithm can work with 84 statistical iris features that are extracted from an individual. Space and time complexity of the proposed approach is lesser than the existing methods. This algorithm has been implemented and results have been analysed on 2500 different iris patterns acquired in India under different real-time conditions. Experimental results illustrate that the proposed method has been easily espoused in elections, bank transactions and other security applications.