Iris feature extraction and recognition based on Empirical mode decomposition

Iris recognition is one of effective methods for the biometrics recognition, while Empirical mode decomposition (EMD) is an effective technique for non-linear, non-stationary signal analysis. In this paper, an Iris recognition method based on an improved feature extraction approach is proposed, in which iris signal is decomposed into several Intrinsic Mode Functions (IMFs) by EMD first, and then one or several IMFs suitable for recognition are taken as a feature vector. The suitable IMFs represent the most important information of iris images, so that they give more contributions to the recognition. The experiments show that the chosen IMFs can reduce noise interference and extract the iris features effectively and the proposed algorithm can achieve an excellent performance.