Iris fusion for multibiometric systems

The widespread interest in personal identification has increased the need for an accurate and efficient ways of identification, verification and authentication. Biometric-based personal identification systems have proven their superior performance. These systems rely on the person physiological traits such as the iris, fingerprint or face, etc. It is worth noting that these systems are more effective than conventional personal identification systems that are based on passwords and/or smartcards. In recent years, there has been emergence in the fusion/combination of multi-biometric traits to further enhance the performance of biometric systems. The latter are commonly known as multi-biometric systems. In this paper, we propose a novel scheme for the fusion of iris images prior to the feature level where we model the iris textures using the Generalized Gaussian Distribution (GGD). Then, a systematic pattern retrieval algorithm is applied in order to improve the accuracy of overall system. Normalized iris sub-images are fused based on a specific quality measure. Simulation results clearly indicate the improvement in performance due to the proposed iris fusion scheme.