An efficient approach towards iris recognition with modular neural network match score fusion

The recognition system when concerned with high-security, authentication of authorized person and invalidation of an imposter is a vital task for the system. The system based on iris feature is considered as highly secure and most reliable system due to the intrinsic property of an iris. This paper presents an efficient iris recognition approach based on the fusion of modular neural network output scores in order to improve the recognition performance of the system. Multimedia University V2 (MMU2) Iris database obtained from the public domain digital repository has been used to evaluate the performance of the proposed approach by considering different performance measures. The experimental results demonstrate the efficiency of the proposed approach for both identification and verification performance of the system. The proposed approach achieved 98.57% recognition accuracy in verification mode and 96.86 % recognition accuracy in identification mode with the considered dataset.