Iris Recognition System Based on Fractal Dimensions Using Improved Box Counting

摘要

In this paper, a new approach based on an improved fractal dimensions (FDs) for an iris recognition system is proposed. Improved FDs, based on a differential box-counting (DBC) method were extracted from horizontal coefficients, vertical coefficients, and diagonal coefficients in different levels, by discrete wavelet transform. The CASIA iris images database was used to evaluate the algorithm; the simulation results are very promising.