Thesis Type: Postgraduate
Institution Of The Thesis: Gazi Üniversitesi, Fen Bilimleri Enstitüsü, Turkey
Approval Date: 2016
Student: BUĞRA ŞİMŞEK
Supervisor: NURSEL AKÇAM
Abstract:In recent years, the biometric systems have acquired a widespread usage in many different areas. These systems ensure the identification and recognition of persons by using unique features such as fingerprint, retina, iris, face shape, and hand vessels. Iris based biometric systems are frequently used systems and have lower false matching rates compared to other biometric systems. Since iris recognition systems are not harmful to human eye and since the structure of human iris changes quickly after the death of the humans, iris recognition systems are used largely in real world applications. In this study, iris comparison algorithm is accelerated by using new hardware and software technologies and it is ensured that it is compatible with the heterogeneous systems. Iris comparison algorithm is parallelized via multi core hardware substructure with synchronous threads. For the implementation, C based OpenCL language is used which maintains a heterogeneous, mobile and parallel computing substructure. Hamming distance algorithm is parallelized and it is ensured that it is compatible with heterogeneous systems by using all the benefits of OpenCL.