Iris recognition model based on Curvelet transform and least square support vector machine

摘要

Iris image quality is not high because of the eyelids, eyelashes, and other interference information. This paper puts forward an automatic iris recognition model based on the discrete Curvelet transform and least squares support vector machine (LSSVM). Firstly, iris image is preprocessed, and then the discrete Curvelet transform is used to decompose the iris image to extract the feature vector in coarse scale and fine scale. Finally, the feature vectors are input to LSSVM to recognize the iris image, and the simulation experiment is carried out on CASIA iris database. The simulation results show that the proposed model can extract the superior characteristics of iris, and improve iris recognition correct rate compared with the other models, it has good application value.