Abstract—Iris is one of the best biometric features for security applications. This paper focuses on the iris recognition and classification system and its performance in biometric identification system. The steps of iris recognition include image normalization, feature extraction and classifier. This work is an application of Patrick Simpson’s fuzzy min-max neural network (FMN) Classification. Fuzzy min-max classification neural networks are built using hyperbox fuzzy sets. We performed comparative studies of different similarity measures applied to various classifiers. The feasibility of the FMN Classification Algorithm has been successfully evaluated on CASIA database with 756 images and found superior in terms of generalization and training time with equivalent testing time.

Index Terms—Biometrics, Feature extraction, Fuzzy Min-max neural network.

S. S. Chowhan, Dept. of Computer Science, COCSIT, Latur, India.(e-mail: csantu_149@rediffmail.com )
G. N. Shinde, Dept. of Electronics and Computer Science, Indira Gandhi College, Nanded, India. (e-mail: shindegn@yahoo.com ).

Cite: S. S. Chowhan and G. N. Shinde, "Iris Recognition Using Fuzzy Min-Max Neural Network," International Journal of Computer and Electrical Engineering vol. 3, no. 5, pp. 743-747, 2011.