A novel iris recognition method based on feature fusion

Although many approaches for iris recognition have been proposed in the last few years, few of them can perfect well in various image qualities. In this paper, a novel method for iris recognition based on feature fusion is presented. Global and local iris features are extracted to improve the robustness of iris recognition for the various image quality. To represent the iris pattern efficiently, the global features are obtained from the 2D log Gabor wavelet filter and the local features are fused to complete the iris recognition. The weighting Euclidean distance and the Hamming distance are applied to match and classify. In addition, the thresholds are set up to reduce the computation time of match, and to increase robust iris recognition. Experimental results that confirm the benefits of using the proposed method are reported.