IRIS Recognition Systems for Biometric Applications

Abstract

Biometrics gives more security compared to digital passwords patterns reducing security concerns in today’s world. Bio-metric Industry is growing in a tremendous manner due to the recent initiative by the government of India to make the digitized identity of Indian Citizens. Iris Recognition considers as more of the secure and reliable biometric authentication system in Industry. This Thesis aims to provide detail explanation of each step of Iris recognition and introducing various algorithms that helps to locate iris in an effective manner. In the later section of the thesis, algorithm performance is compared with Libor Masek method with different circle detectors like hough and daughman. In order to show a performance of the proposed algorithm, it is applied on various databases that are publicly available. The Accuracy of Proposed Algorithm was 96.14% in IIT Delhi Iris database 80.14% in CASIA lastly 100% in Mantra Iris Sensor Images. The iris codes of two different eye images are also compared by using 1-D log Gabor filter to show the practical approach of matching images.