Iris classification algorithm based on stable features

Iris matching similarity will be affected by some fluky and unfixed features among iris feature templates, that make iris classification false rate (include false accept rate and false reject rate) increase. In order to resolve this problem, a stable feature extraction algorithm is proposed in this paper, the stable features of an iris can be extracted from its several images, then iris feature template can be built based on these stable features, thus, iris feature template can be used for iris classification. Simulation results show that the selected stable features achieve quite high classification accuracy and the proposed algorithm is effective for improving iris recognition performance on the CASIA Iris database.